Menu
Book a demo
FDA Warning Letters to Food Plants: A Monthly Tracker by 21 CFR 117 Citation

FDA Warning Letters to Food Plants: A Monthly Tracker by 21 CFR 117 Citation

Monthly tracker of FDA warning letters to food and beverage plants, sorted by 21 CFR 117 citation, with the record that answers each one.
FDA Warning Letters to Food Plants: A Monthly Tracker by 21 CFR 117 Citation

In short: An FDA warning letter to a food plant lists each violation against a section of 21 CFR part 117, the preventive controls and good manufacturing practice rule, and asks for a written reply within 15 working days. This tracker sorts each month's letters to food and beverage manufacturers by the section cited, and reads each citation for the record that answers it: the dated line record, or the food safety plan, the lab and the sanitation file.

  • Eight letters issued to food and beverage manufacturers from 1 August to 15 October 2026 were posted by 17 November 2026. Seven cite the hazard analysis (117.130), five the process controls for time, temperature or pH (117.135(c)(1)) and four the sanitation controls (117.135(c)(3)).
  • Three of the eight went to noodle plants, and all three were cited for how cooked noodles cooled. Five letters fault a record that was missing, undated or never reviewed.
  • Example plant: one failed cooling check puts 7,140 lb of fresh rice noodles on hold when the line record shows which batches ran since the last good check, and 51,795 lb when only the production date is known. At the BLS shelf price of US$1.300 to US$1.384 a pound, that is US$9,282 to US$9,881.76 against up to US$71,684.28.

Tracker, updated 18 November 2026. Next update: December 2026, with the letters posted since. Sources are listed at the end.

A food plant answers a warning letter with records, and plants that measure OEE for food and beverage by machine, line and shift already keep one of the two kinds that carry the reply. The food safety plan, the lab results and the sanitation file show what the plant decided and tested. The line record shows what actually ran: when each product ran on which line, when it stopped, when it was cleaned and changed over, and how much it made. A plant with that record can date the food affected by a failed control instead of holding a whole day. This page logs the FDA's letters to food and beverage manufacturers each month, sorts the citations by section, and marks which ones a dated line record helps answer and which ones it cannot.

The scope is human food made in a plant: cheese, noodles, tofu, fermented vegetables, ice cream, baby food and seafood. Letters to importers under the Foreign Supplier Verification Program, to dietary supplement makers, to animal food plants and to drug or device makers are counted below but not sorted.

Food plant warning letters, August to November 2026, by cited section

The rows come from the FDA warning letter listing, searched on 18 November 2026 for the food subjects it uses (insanitary conditions, hazard analysis and risk-based preventive controls, manufacturing, packing or holding human food, seafood HACCP, juice HACCP, acidified foods) and read letter by letter. The period is letters issued from 1 August 2026 and posted by 17 November 2026. The listing posts a letter weeks after it is issued, so a letter dated in late October or November can still appear in next month's edition.

Letter dateFirm and cityProductSections citedInspection endedReply to Form 483
15 Oct 2026CNC Noodle Corporation, Hayward, CAWheat flour noodles, wonton wrappers117.130(a)(1), 117.135(c)(1), 117.135(c)(2), 117.35(a), 117.80(c)(2)28 Jan 2026None received
5 Oct 2026ABC Food Corp, Boston, MARice noodles, tofu117.130(a)(1), 117.135(c)(1), (c)(2), (c)(3), 117.160, 117.20(b)(4)7 Apr 202628 Apr 2026
29 Sep 2026IF Co-Pack, LLC dba Initiative Foods, Sanger, CAFruit and vegetable baby food117.130(a)(1), 117.135(a)(1) and (c)9 Apr 2026None received
29 Sep 2026Dreyer's Grand Ice Cream, Inc., Bakersfield, CAIce cream cone products117.130(a)(1), 117.10(b)20 Feb 202613 Mar and 13 Apr 2026
22 Sep 2026Rama Food Manufacture Corp., Ontario, CA (two sites)Rice noodles117.135(c)(1), (c)(3), 117.20(b)(4), 117.126(a)(1), 117.35(a), 117.35(c)26 Feb 202617 Mar 2026
11 Sep 2026Ocean Pride Seafood, Inc., Ventura, CARaw and smoked fish, shellfishPart 123: 123.6(a), (b), (b)(2), 123.7(b), 123.11(b), with Part 117 subpart B1 Apr 20261 Apr and 6 May 2026
11 Sep 2026RDM Express Inc., San Francisco, CAFermented cabbage117.130(a)(1), 117.135(c)(1), (c)(3), 117.165(a)(4), 117.405(a)(1), 117.41010 Mar 2026None received
28 Aug 2026Tipical Latin Food, Corp. DBA La Pradera, Doral, FLVenezuelan style cheeses117.126(a)(1), 117.130, 117.135(c)(1), (c)(2), (c)(3), 117.165, 117.35(a), 117.35(c), 117.40(a)(1)27 Apr 2026None received

Source: the eight FDA letters linked in the case notes below. Sections are those each letter names for the plant; misbranding citations under the labeling rules in 21 CFR part 101 are left out of the column.

The same eight letters sorted by the part of the rule cited, with the record that has to answer each citation:

What the letter citesSectionLettersRecord that answers itDoes a dated line record help?
Hazard analysis missed a hazard117.1307Food safety plan, written by a preventive controls qualified individualNo: the analysis is a written decision
Process controls: cooling, holding temperature, pH117.135(c)(1)5Monitoring records (117.145) and validationYes, in part: when each batch was cooked, cut, packed and stopped, so the food behind a failed check can be dated
Sanitation controls117.135(c)(3)4Sanitation records, environmental monitoring resultsYes, in part: when each clean took place against the production runs between cleans
Allergen controls117.135(c)(2)3Label review, allergen procedureYes, in part: which product ran on which line after which changeover
Sanitary operations, cleaning of food contact surfaces117.35(a)3Cleaning SOP and sanitizer checksOnly the time of the clean; not the sanitizer strength
Condensate dripping onto food117.20(b)(4), 117.80(c)(2)3Maintenance and building recordsOnly the time a line stopped for the repair, if the stop was logged
Supply-chain programSubpart G, 117.405, 117.4103Approved suppliers, supplier verificationNo: ingredient lots and genealogy live in the ERP or traceability system
Verification and records review117.1653QA review of records within 7 working daysNo
No food safety plan at all117.126(a)(1)2Food safety planNo
Pest exclusion117.35(c)2Pest control recordsNo
Validation of a lethal step117.1601Validation studyOnly the time each batch spent in the step
Ingredient specification and certificate review117.135(a)(1) and (c)1Receiving and supplier recordsNo
Equipment design and upkeep117.40(a)(1)1Maintenance recordsNo
Personal hygiene, gloves, hand washing117.10(b)1Training and supervision recordsNo
Seafood HACCP plan, monitoring and sanitation monitoringPart 1231HACCP plan and CCP recordsYes, in part: receiving and storage times by product

The sort rule, printed so the next edition uses the same one: a letter counts once in a row when it names the section for this plant, either as a numbered violation or as a missing element of the food safety plan. Hand hygiene faults written under a sanitation control count in the sanitation row, not under 117.10(b). Seafood plants are held to Part 123, with the good manufacturing practice in Part 117 subpart B.

Three readings carry the rest of the page.

  • The plan is cited more than the plant. Seven of eight letters fault the hazard analysis, and two plants had no food safety plan at one site. No record of the line repairs a hazard analysis that left out Listeria, B. cereus or patulin.
  • Time and temperature is where the line record comes in. All five process control citations concern how long food stayed warm, cooled, or sat before a pH check. Each one is a question about when, and the answer is a timestamp.
  • Missing records are cited as often as missing controls. By this page's count, five letters fault a record: undated pH logs and sanitation logs that name no equipment (RDM Express), no receiving or cooler records since 2022 (Ocean Pride), an undocumented sanitizer check (ABC Food), no documented review of supplier certificates (Initiative Foods) and undocumented root cause work on Listeria positives (Dreyer's).
What the food letters cite8 FDA warning letters, 1 Aug to 15 Oct 2026Hazard analysis, 117.1307Process controls, 117.135(c)(1)5Sanitation controls, 117.135(c)(3)4Allergen controls, 117.135(c)(2)3Sanitary operations, 117.35(a)3Condensate, 117.20 and 117.803From inspection to letter183days, medianto the letter4 of 8no replyto the 4835 of 8fault arecordThe plan says what should happen.The line record says when it ran.
Eight FDA warning letters to food and beverage manufacturers, issued 1 August to 15 October 2026, counted by the 21 CFR section each cites. See the FDA listing; the sort rule is printed above.

From Form 483 to warning letter: the 15 working days

A warning letter is the second written step after an inspection. At the close of an inspection, the investigator hands management a Form FDA 483, a notice of “objectionable conditions”. The FDA's Form 483 FAQ says the form does not constitute a final Agency determination of a violation; the agency weighs it with the inspection report, the evidence collected and the firm's reply. The FAQ encourages a written reply with a corrective action plan, and gives no deadline.

The warning letter follows when the FDA believes it has found significant violations (About Warning and Close-Out Letters). Each of the eight letters here asks for a written reply within fifteen working days of receipt, stating the corrections, the steps that stop a recurrence and the documents that show both; where the work takes longer, the plant states the reason and the date it will be done. Each letter warns that a failure to correct may lead to legal action, such as seizure or injunction.

StepWhat happensWhat the plant needs from its records
1. InspectionInvestigators observe production, take samples and swabs, and copy recordsRecords made at the time of each activity, with the date and, where needed, the time
2. Form FDA 483Observations handed over at the close; not a final determinationA written reply with a corrective action plan
3. Warning letterSignificant violations named by section; reply due in 15 working daysCorrections, the cause, and records that show the fix running
4. Close-out letterIssued only after corrections are made and verified by the FDAThe same records at the follow-up inspection

The gap between steps 2 and 3 is long. In the eight letters, the letter came 123 to 260 days after the inspection ended, with a median of 183 days. Four of the eight plants had sent no reply to the Form 483 by the date of the letter. Of the four that did reply, the letters name what the reply left out: Rama's revised plan still allowed 3.5 hours above 70°F, Ocean Pride's revised HACCP plans were “inadequate”, ABC Food sent no documents after its first reply, and Dreyer's did not say why earlier corrective actions failed.

A close-out letter needs more than a promise. The FDA's page says the corrective actions “must actually have been made and verified by FDA”, and that verification normally takes a follow-up inspection. Four of the eight letters (Rama, Dreyer's, ABC Food, CNC Noodle) say outright that the fixes will be checked at the next or a follow-up inspection.

That next inspection may be years away. The Food Safety Modernization Act sets a target of one inspection every 3 years for each high-risk domestic food facility and every 5 years for the others. GAO-25-107571, released on 8 January 2025, found the FDA has not met its targets since fiscal year 2018 and averaged 8,353 domestic food inspections a year from fiscal 2018 to 2023. The reply goes out in weeks; the verification can come years later, and the records have to cover the whole gap. Under 21 U.S.C. 350g(g), the statute behind Part 117, the records of monitoring, corrective actions and testing are kept for not less than 2 years.

What a record has to show under 21 CFR 117

Part 117 sets the standard every citation is measured against, and the three sections below decide whether a record can answer a letter.

  • The record is made at the time. 21 CFR 117.305 requires records to be “created concurrently with performance of the activity documented”, with the actual values observed, the plant's name, the person who did it, the product and lot code where appropriate, and “the date and, when appropriate, the time” of the activity. A pH log without times, as at RDM Express, falls short of that last line.
  • Monitoring has a stated frequency. 21 CFR 117.145 asks for written monitoring procedures, including how often each check is done, at a frequency that shows the control is consistently performed. For refrigerated food, either affirmative records of temperature control or exception records of its loss are allowed.
  • Affected food is evaluated and held. 21 CFR 117.150 requires corrective action procedures that identify and correct the problem, reduce the chance it recurs, evaluate all affected food for safety and keep it out of commerce if it cannot be shown to be safe. Corrective actions are documented in records.

The third point decides the size of every hold. “All affected food” is the food made between the last check that passed and the check that failed. A plant that knows to the minute when each batch ran between those two checks holds those batches. A plant that knows only the production date holds the day. The worked example below prices the difference.

Case notes, newest first

Each note ends with what its own letter tells a plant to check. New letters are added at the top, with the letter date and the FDA's case number.

CNC Noodle Corporation, 15 October 2026 (CMS 732000): wheat not declared, cooling not evaluated

The CNC Noodle letter follows an inspection from 13 to 28 January 2026 of a Hayward, California plant making pan fried, egg, steam and chao mian noodles and wonton wrappers. The hazard analysis did not evaluate B. cereus and C. botulinum in pan fried noodles packed in heat sealed pouches, which get no lethal step and no pH or water activity control; investigators recorded the cooling of one lot on 20 January. The analysis also left out undeclared allergens although every fresh noodle contains wheat, and wheat was missing from labels on several products. The firm recalled steam and chao mian noodles for undeclared wheat on 2 March 2026. Food contact surfaces were cleaned with water only on 13 January and by scraping and compressed air on 15 January, a dirty fan blew onto cooling noodles and condensate dripped onto the steamed noodle line.

  • A cooling check needs the cooking time beside it. The letter rests on temperatures taken while one lot cooled. A plant can only show cooling met a limit if it can show when that lot came off the cooker.
  • An allergen on every product is still an allergen control. All products contain wheat, so the risk sits on the label, and the label review has to be in the plan. The line record cannot answer this citation.

ABC Food Corp, 5 October 2026 (CMS 730984): no hazard analysis for any food

The ABC Food letter covers a Boston plant making ready-to-eat rice noodles and tofu, with an FDA inspection from 17 March to 7 April 2026. The firm had not conducted or documented a hazard analysis for any food it makes. On 17 March, batches of Ho Foon cut noodles in reduced oxygen packaging cooled in the freezer from 76°F to 62 to 66°F over about 5 hours. On 19 March, noodle mixture for 2 lb bags entered the lethal treatment step at 57°F and left after 56 seconds at 149°F, with no validation (117.160). The production manager measured sanitizer at about 100 ppm with a strip on 18 March without recording it on the form the firm's own program requires, and noodles were made on that equipment on 19 March. Condensate dripped from an overhead water pipe onto packaging for a silken tofu lot.

  • A 56 second step needs its time on record. A lethal step that is a matter of seconds at a set temperature is validated and monitored by time. Both numbers have to sit in the same record.
  • The sanitizer check was done, but not written down. The reading existed for a moment on a strip. Under 117.305 a record made at the time is what counts.

Initiative Foods, 29 September 2026 (CMS 732244): patulin and a certificate nobody reviewed

The Initiative Foods letter covers a Sanger, California co-packer of high-acid, acidified and low-acid canned fruit and vegetable baby food, with an FDA inspection from 31 March to 9 April 2026. The hazard analysis for pouches rated patulin as low likelihood and set no patulin specification for incoming ingredients. Some certificates of analysis carried patulin results, but there was no documentation that they were reviewed. On 26 December 2025 the plant used an ingredient lot whose certificate had no patulin result. The letter cites the firm's February 2026 recall of an apple pear banana puree, after an FDA Total Diet Study sample showed elevated patulin and the firm's own test of retained samples found 40 ppb in one lot.

  • The receiving record is where this one is answered. Each ingredient lot, its certificate and the review belong in the supply-chain file, and the ingredient to finished lot link lives in the ERP or traceability system. A line record adds only the date the product ran.
  • A missing result is a result. A certificate without the one value the plan relies on is a reason to hold the ingredient lot before 26 December, not after a recall.

Dreyer's Grand Ice Cream, 29 September 2026 (CMS 727644): 19 Listeria positives and no documented root cause

The Dreyer's letter follows an inspection from 27 January to 20 February 2026 of a Bakersfield, California plant making frozen ready-to-eat cone products. The hazard analysis treated Listeria and Salmonella as not significant, relying on a sanitation prerequisite program, although product is exposed and handled before packaging and gets no lethal step after it. Environmental monitoring from 2021 to 5 February 2026 showed recurring findings; in 2026 to that date there were 19 Listeria positives, 8 of them confirmed L. monocytogenes. The firm's own procedure called for root cause analysis and documented corrective action on repeat positives, which in practice were not documented. The sampling plan had no written rationale tying it to production volume, and glove tests measured Enterobacteriaceae, which does not show Listeria. Investigators saw glove faults on four products on 28 January, 11 February and 13 February.

  • A positive needs what ran that day. A root cause on a repeat positive starts with the line, the product, the cleans and the stops around the swab date. That part the line record holds; the swab result is the lab's.
  • Sample by volume, and write down why. The FDA could not judge whether the sampling plan matched production because no rationale was written. Run hours and output by line are the volume.

Rama Food Manufacture, 22 September 2026 (CMS 730689): noodle sheets above 70°F for 6 hours and 10 minutes

The Rama letter covers two Ontario, California sites making ready-to-eat rice noodles, with an FDA inspection from 27 January to 26 February 2026. On 12 February, cooked noodle sheets stayed above 70°F for 6 hours and 10 minutes; the firm had not named cooling as a preventive control and did not monitor it. The letter cites the Food Code's cooling limits, from 135°F to 70°F within 2 hours and to 41°F within 6 hours in total, and a limit of three hours in total above 70°F for rice. Sheets cooked on 12 February were cut and packed at about 05:30 on 13 February, and one bag read 61°F at 09:13. Lot 04-23-A1 cooled only 12°F in 26 hours of refrigeration, and lot 04-23-B1 only 4.5°F. FDA samples of three products read pH 4.55 to 4.91, above the firm's own target, with water activity of 0.98 to 1.00. Cut machines, trays and tables used daily were cleaned on a rotating basis; swabs found Listeria grayi and Listeria innocua; condensate dripped onto exposed sheets on 27 January; the second site had no food safety plan and no written recall plan.

  • Every number in this letter is a time. Cooked at, cut at, packed at, read at. A line record with the time each batch left the steamer and each pack run started gives the cooling clock its start.
  • Rotating cleans need a count. Equipment used every day but cleaned in rotation means some units run several days between cleans. Counting production days per unit between cleans turns that into a number the plan can limit.

Ocean Pride Seafood, 11 September 2026 (CMS 730154): no cooler records since 2022

The Ocean Pride letter covers a Ventura, California seafood processor of raw tuna, salmon and other fish for raw consumption, smoked salmon, imitation crab and shellfish, with an FDA inspection from 23 March to 1 April 2026. It is a Part 123 seafood HACCP letter, with Part 117 subpart B for sanitation. Monitoring listed in the HACCP plan was not done at receiving and storage, a repeat of the November 2019 inspection: there were no receiving records since 23 November 2022, no cooler temperature records since 31 December 2022 and no sanitation monitoring records since 21 August 2023. One histamine plan covered ten species, six of which carry a parasite hazard instead. On 23 March an employee washed a gutted salmon on the floor over a trench drain, then filleted it.

  • Three years of silence is one finding. A plan with critical control points and no records for the period cannot be shown to have run at all. The reply asks for at least five consecutive production days of records.
  • Plans by species, records by species. Splitting one plan into species plans means the receiving and storage records have to name the species, with the time it came in.

RDM Express, 11 September 2026 (CMS 729223): a pH log with no times

The RDM Express letter covers a San Francisco maker of refrigerated fermented cabbage, with an FDA inspection from 18 February to 10 March 2026. The food safety plan of 30 April 2025 relied on a finished product pH to control C. botulinum and L. monocytogenes; an FDA sample on 25 February 2026 read pH 4.0 to 4.17 with water activity 0.98. pH was taken from one spot in a large barrel after fermentation and not at packing, which took place days later, and the daily pH log records no times. Sanitation logs did not name the equipment cleaned or who cleaned it, records were not reviewed within seven working days, there was no environmental monitoring, and cranberries used without a lethal step had no supply-chain control.

  • A pH value without a time is not a monitoring record. The letter asks when the check happened against fermentation, cold storage and packing. Only a time answers that.
  • Packing days after fermentation is a gap to measure. The days between fermentation and packing are a number from the production record. The plan's pH check has to cover the end of that gap, not its start.

Tipical Latin Food (La Pradera), 28 August 2026 (CMS 730405): no food safety plan, cited a third time

The La Pradera letter covers a Doral, Florida maker of ready-to-eat queso de mano, queso guayanes, queso estilo telita and cachapas, with an FDA inspection from 20 to 27 April 2026. The firm had no food safety plan, a repeat of the inspections of December 2024 and May 2025. Soft cheese was held overnight in a walk-in cooler with no temperature monitoring. Employees put on gloves without washing hands on 20 and 21 April, cheese residue was found on molding machine screws said to be clean, and white residue on molds after sanitizing. Flies landed on product and food contact surfaces, about seven cockroaches sat on a box of curd bags on 21 April, and a door to the outside stood open, the pest finding repeated from December 2024. Milk, the allergen in the cheeses, had no labeling control.

  • Three inspections, one missing document. The plan itself is the reply here. No production data can stand in for a hazard analysis that was never written.
  • Overnight in the cooler is a time and a temperature. The time the last batch went in and the time it shipped bound the hold; the cooler log supplies the temperature.

Letters in the period not sorted above: importers under seafood HACCP (South Asian Food, 20 August; Star Ocean Wholesale Foods, 1 September), nine Foreign Supplier Verification Program letters to importers from 5 August to 27 October, one dietary supplement letter (19 October) and three animal food letters (5 and 21 August). Letters posted in the period but issued before August are left to their own month: La Cabana Food Distributing (29 July), H2 BEV and H2 RENU (31 July), and two from March and April.

Case: three noodle plants, one cooling question, October 2026

Three of the eight letters went to noodle makers in four weeks: Rama on 22 September, ABC Food on 5 October and CNC Noodle on 15 October. They make different noodles, rice sheets in Ontario, Ho Foon cut in Boston, wheat noodles in Hayward, and the same hazard runs through all three letters: spore forming bacteria in a cooked starch product that is not acidified enough and not cooled fast enough.

PlantWhat the letter measuredLimit the letter points toThe time that decided it
Rama, rice sheetsAbove 70°F for 6 h 10 min on 12 Feb; lots cooled 12°F and 4.5°F in 26 h135°F to 70°F in 2 h, to 41°F in 6 h; 3 h total above 70°F for riceCooked, cut at about 05:30, packed, read at 09:13
ABC Food, Ho Foon cut76°F to 62 to 66°F over about 5 h in the freezer on 17 MarNo limit set; no preventive controlStart and end of the freezer cooling
CNC Noodle, pan friedCooling of one lot recorded by investigators on 20 JanB. cereus and C. botulinum not evaluatedWhen the lot left the fryer and when it was sealed

Source: the three FDA letters. In all three the gap was the same: no record of when the cooling clock started. Rama's figures exist because FDA investigators took the temperatures. ABC Food cooled in a freezer, and its reply speaks of cooling solutions within limited freezer space. CNC's lot was timed by the investigators, not by the plant.

What this case changes for a plant making cooked noodles, rice or pasta:

  • Start the cooling clock from the line, not from the cooler. The batch leaves the cooker or steamer at a time the line record already holds. The cooling check measures against that time.
  • Check pH at packing, from more than one spot. Rama's and RDM's letters both fault a pH sample taken from one place, at one time, before packing.
  • Do not load warm product into reduced oxygen packaging. Rama's 10 lb bags were packed with the air pressed out while still warm, and two letters name reduced oxygen packaging as the condition for C. botulinum.

Book a demo and see OEE and losses per line and shift.

Worked example: one failed cooling check on a fresh rice noodle line

Example plant. Every operating figure in this section is Example data; the cooling limits, the hold rule and the price per pound are sourced. The plant is not any firm named above.

The Example plant makes fresh rice noodles on two lines, steamed in sheets, cooled on carts, cut and packed in 10 lb bags. Its lot code is the production date and the line. Its food safety plan names cooling as a process control, with the limits the Rama letter cites: 135°F to 70°F within 2 hours. A cart from each line is probed at 06:15, 09:45 and 13:15, and the reading goes on the cooling log with the time.

On Thursday 15 October 2026 the 09:45 check on line 2 passes. The 13:15 check on line 2 finds a cart from batch 2-08 at 78°F, 2 hours 54 minutes after it left the steamer. The critical limit is missed. Under 117.150 the plant has to evaluate all affected food and keep it out of commerce until it is shown to be safe. The question is which food that is.

Step 1: the hold from the line record. The line record shows each batch on line 2, the time it left the steamer and the pounds made from it. It also shows a 34 minute stop at 10:52, which the operator logged at the machine with the reason “blast cooler door”. The stop tells the investigation where to look first; it does not shorten the hold, because the 09:45 check is the last proof that cooling was in control.

Line 2 batch, 15 OctLeft the steamerMade (lb)
2-0709:521,184
2-0810:211,207
Stop, “blast cooler door”10:52 to 11:260
2-0911:291,163
2-1011:581,219
2-1112:271,176
2-1212:561,191
Hold A: line 2, 09:45 to 13:157,140

Example data. 1,184 + 1,207 + 1,163 + 1,219 + 1,176 + 1,191 = 7,140 lb. Batch 2-06 left the steamer before the 09:45 check and is outside the window.

Step 2: the hold from the lot code. Without the times, the smallest unit the plant can name is its lot code: line 2, 15 October. Line 2 made 26,418 lb that day across two shifts (Example data), and all of it is held.

Step 3: the hold from the production date. If the line is missing from the record too, the plant holds both lines' output for the date: 26,418 + 25,377 = 51,795 lb (Example data).

Step 4: price each hold. No federal series prices fresh rice noodles. The nearest published price per pound is the BLS average price of spaghetti and macaroni, US city average, which ran from US$1.300 (November 2025) to US$1.384 (October 2026) over the twelve months published to date. It is a shelf price for dry pasta, so it values a pound of noodles at retail; read the figures as the scale of the hold, not as the plant's invoice. If the held food cannot be shown safe and is destroyed:

HoldDrawn fromlbAt US$1.300 a lbAt US$1.384 a lb
ALine record: line 2 batches since the last good check7,140US$9,282.00US$9,881.76
BLot code: line 2, whole day26,418US$34,343.40US$36,562.51
CProduction date: both lines51,795US$67,333.50US$71,684.28
B minus ACost of having no batch times19,278US$25,061.40US$26,680.75
C minus ACost of having no line in the lot code44,655US$58,051.50US$61,802.52

Example pounds; price band from BLS series APU0000701322. 7,140 x 1.300 = 9,282.00; 7,140 x 1.384 = 9,881.76. 26,418 x 1.300 = 34,343.40; 26,418 x 1.384 = 36,562.51. 51,795 x 1.300 = 67,333.50; 51,795 x 1.384 = 71,684.28. 26,418 minus 7,140 = 19,278; 51,795 minus 7,140 = 44,655.

One failed check, three holdsExample plant, fresh rice noodles, lb on holdA. Line record: 7,140 lbB. Lot code, line 2: 26,418 lbC. Date, both lines: 51,795 lbAt the BLS price of US$1.300 to US$1.384 a lb:A: US$9,282 to US$9,882B: US$34,343 to US$36,563C: US$67,334 to US$71,684
Example data. Bars to scale, about 9.07 px per thousand lb. Price band from the BLS average price of spaghetti and macaroni, November 2025 to October 2026.

Step 5: price the containment stop. Line 2 stops from 13:15 to 14:05 while carts are segregated and the cooler is checked: 50 minutes. Running, line 2 makes 41 lb a minute (Example data), so the stop costs 50 x 41 = 2,050 lb of output, or 2,050 x 1.300 = US$2,665.00 to 2,050 x 1.384 = US$2,837.20 at the same price band. The stop sits in line 2's availability for the shift whichever hold the plant can draw.

Step 6: read the result. The failed check is the same in all three holds; only the record differs. With batch times the plant holds 7,140 lb, worth US$9,282.00 to US$9,881.76 at shelf price. With only a lot code by line it holds 26,418 lb, and with only a date 51,795 lb, up to US$71,684.28. The difference between A and C, US$58,051.50 to US$61,802.52, is the price of not knowing which batches ran when. The same record also answers the warning letter question the three noodle plants could not: when did the cooling clock start.

Three caveats. A shelf price is above a plant's selling price for dry pasta and is used here only for scale. Testing and releasing held food, labour to segregate and rework, disposal, and any recall cost are not priced. And the hold from the line record is only as good as the cooling log: without a passing check at 09:45 with its time, hold A has no start.

Plan, lab, sanitation file or line record: what to fix first

The eight letters point to five places a plant can put its next month of effort. They are not alternatives for a plant with no plan; a food safety plan comes first, and two of the eight letters went to plants without one.

FixChoose it first ifAnchor on this pageWatch for
Write or redo the hazard analysisThere is no plan, or the plan has not been reanalyzed since a positive, a recall or a new product7 of 8 letters cite 117.130; La Pradera cited at three inspectionsA plan written by a consultant that no one on the line has read
Set and validate process control limitsCooked food cools, rests or ferments before packing5 letters on 117.135(c)(1); Rama's 6 h 10 min above 70°FLimits copied from a guide without a challenge study for the product
Environmental monitoring with written rationaleReady-to-eat product is exposed after the lethal stepDreyer's 19 positives in five weeks of 2026; RDM with no programRepeat positives closed without a documented root cause
Supplier verification and certificate reviewAn ingredient carries a hazard the plant does not kill, such as patulin in fruit or parasites on berriesInitiative Foods' 40 ppb lot and February recall; RDM's cranberriesCertificates filed but never compared with a specification
Dated line record joined to the monitoring logsThe plan exists, but holds are drawn by date because batch times are missingExample hold A against C: US$58,051.50 to US$61,802.52A record that misses stops, changeovers and cleans, or a clock that drifts from the cooling log

Questions to put in writing to a supplier of line data or food safety software:

  • Is each batch, product run, stop, clean and changeover recorded with its line and its time, at the time it happens?
  • Can the output of one product on one line be summed between two timestamps, such as two monitoring checks?
  • Can a stop be given a reason at the machine, and can the reasons be read for one line and one day?
  • How do the line record's times join to the monitoring log, the sanitation log and the lot codes in the ERP or traceability system?
  • Can the record show the production days each piece of equipment ran between two cleans?

The trade-offs between the two kinds of record are in HACCP records vs OEE data, and how to count time lost to food safety checks without hiding it is in food safety stops in OEE.

A 15 working day reply and a quarter of fixes, by role

A warning letter, a Form 483 or a near miss on a cooling check all lead to the same list. Each step produces a record the next inspection can read.

This week: plant and quality manager

  1. Read each citation in the letter or the Form 483 and write next to it which record answers it: the plan, a monitoring log, the lab, the sanitation file, pest control or the line record.
  2. For each process control, check that the monitoring log records the time of each check, not only the date and the value, as 117.305 requires.
  3. Pick last week's busiest day on one line and try to list, from records alone, which batches ran between two monitoring checks. Time how long it takes.

This week: production manager

  1. Make sure the line log holds stops, cleans and changeovers with their times and a reason given at the machine, including stops for food safety checks and repairs to drips or doors.
  2. Check that the lot code names the line as well as the date. If it does not, a failed check on one line holds the other.

This quarter: sanitation and maintenance leads

  1. For equipment cleaned in rotation, count the production days between cleans for each unit from the line record, and set a maximum in the sanitation plan. The master sanitation schedule is where that limit lives, and ATP swab testing verifies the clean itself.
  2. List every overhead pipe, vent hood and fan above exposed product, as three of the letters faulted, and log each fix as a stop with its time. Drips and dirty fans are also a route for foreign material, the subject of the foreign material recalls tracker.

This quarter: the food safety team

  1. Run a mock hold: pick a past monitoring check, pretend it failed, and draw holds A, B and C as in the example. Price them at the product's selling price.
  2. Write the environmental monitoring rationale against production volume, using run hours and output by line, as the Dreyer's letter asks.
  3. Check allergen changeovers against the label review: which product ran on which line after which clean. Allergen changeover validation covers the test that proves the clean worked.

The line record in a food safety file

A food safety plan says what should happen; the monitoring logs say what a check found; the line record says what ran, when, and how much. Of the citations in this month's letters, the plan, the lab and the sanitation file answer the larger share. The line record answers the questions of time: when a batch left the cooker, how long it sat before packing, what ran between two checks and between two cleans.

Fabrico is a cloud OEE and light MES platform. It shows OEE and losses by machine, line and shift, with data from the machine's PLC or a Fabrico IoT device, and operators give the reason for a stop at the machine. With the selling price and margin of each SKU entered, losses are priced in money. The line record shows when each product ran on which line, how many units it made, and its stops and changeovers, cleans included, so the output between two monitoring checks can be read and counted. On manual stations, AI cameras (an add-on) count output. The food safety plan, the monitoring and lab records, and lot genealogy stay in the food safety, ERP or traceability systems, and they join to the line record on the line and the time. Food plants can see how this applies in OEE for food and beverage and in tracking downtime on food and beverage lines; the building blocks are OEE software and a light MES.

Book a demo and see OEE and losses per line and shift.

What people ask next

How long does a food plant have to answer an FDA warning letter?

Each of the eight food letters in this tracker asks for a written reply within fifteen working days of receipt. The reply states the corrections, the steps taken to stop a recurrence and the documents that show both. Where a correction takes longer, the plant gives the reason and the date it will be complete.

What is the difference between a Form 483 and a warning letter?

A Form FDA 483 lists objectionable conditions at the close of an inspection, and the FDA says it is not a final Agency determination. A warning letter follows when the FDA believes it has found significant violations, names them by section, and warns of legal action such as seizure or injunction if they are not corrected. In this tracker the letters came 123 to 260 days after the inspection ended.

Which 21 CFR 117 sections do FDA warning letters to food plants cite?

In the eight letters to food and beverage manufacturers issued from 1 August to 15 October 2026, 7 cite the hazard analysis (117.130), 5 the process controls for time, temperature or pH (117.135(c)(1)), 4 the sanitation controls (117.135(c)(3)) and 3 each the allergen controls, sanitary operations (117.35(a)), condensate, the supply-chain program and verification. Two plants had no food safety plan at one site.

How does a food plant get a warning letter closed out?

The FDA issues a close-out letter only after the corrective actions have actually been made and verified by the FDA, normally at a follow-up inspection. A promise of future action is not enough. Four of the eight letters in this tracker say the fixes will be assessed at the next or a follow-up inspection.

What must a food safety record show under 21 CFR 117?

Under 21 CFR 117.305 a record is made at the time of the activity and shows the actual values observed, the plant, the person who did it, the product and lot code where appropriate, and the date and, when appropriate, the time. Under 21 U.S.C. 350g(g) the records of monitoring, corrective actions and testing are kept for not less than 2 years.

How much food is on hold after a failed cooling check?

All affected food, which under 21 CFR 117.150 is evaluated for safety and kept out of commerce until it is shown safe. In the Example plant on this page, a failed check holds 7,140 lb when batch times show what ran since the last good check, 26,418 lb by a line and date lot code, and 51,795 lb by date alone.

Sources

All sources accessed on 18 November 2026. Dates are letter, release or publication dates where a page gives one; other dates are labelled as what they are.

Track OEE and downtime in real time
See where lines lose time
Book a live demo.
See where lines lose time
See where lines lose time

Real Fabrico tiles, sample data.

By clicking the Accept button, you are giving your consent to the use of cookies when accessing this website and utilizing our services. To learn more about how cookies are used and managed, please refer to our Privacy Policy and Cookies Declaration