Manaz Afzal Book a consultation

Work snippet

n8n and Make error monitoring with retries and a dead-letter queue

Concept DesignIllustrative integration walkthrough. Business details, figures and records are worked examples.

A five-branch service group runs fourteen automations between n8n and Make. When one fails, the error is sorted by type, retried automatically only if a repeat cannot double-post, otherwise set aside with its full payload, and sent to the person who owns it at the right level of urgency.

The problemAutomations fail silently, and retries sometimes create duplicates.

What changesEach failure is classified, retried safely or parked, and reported to someone who can fix it.

Area
Automation reliability
Tools
n8n · Make · Airtable · Slack
Typical scope
US$1,500-3,000
Timeline
About 4 weeks · fixed milestones

The problem

Silent failures leave invoices unposted, and no one knows which run broke or whether re-running it will post everything twice.

  • The same failure is retried in a tight loop and hits the rate limit harder
  • A partial write leaves a job in one system and not in the other
  • Alerts fire 20 times for one outage, then get muted
  • Nobody owns the item, and the payload is already gone

What discovery would examine

A read-only review of 14 Make scenarios and n8n workflows for a five-branch service group, plus two weeks of run history and a 45-minute call with the operations lead.

Findings from the review
IDFindingEvidenceRisk if left
F1Three invoice and bill flows retry in a tight loop on a 429Run history: 40-90 repeats in under two minutesRate-limit lockout; delayed postings
F2Two flows have no error path; a failed run leaves no payloadScenario export: no error handler routeThe item cannot be replayed
F3Slack receives one message per failed item, with no groupingA Tuesday outage produced 86 messages in 11 minutesThe channel is muted, then silence
F4A job-create flow is not idempotent on replayA manual rerun created a second ServiceM8 jobDuplicate jobs and double billing risk
F5No heartbeat on six overnight flowsA paused Make scheduling switch was noticed at 09:40A whole night of work is missed
F6Credentials for two connections expire with no warningVendor email sat in a personal inboxAn outage that looks like a mapping bug

Approach

  1. List every automation

    Owner, trigger, destination and what a failure costs, written in one inventory.

  2. Write the rules

    Which errors retry, how often, when to park, and who is alerted.

  3. Build one handler

    A shared pattern for Make and for n8n, plus an Airtable queue and heartbeats.

  4. Roll out and tune

    One automation at a time, with a rollback that leaves the original steps in place.

Current state and future state

How work moves today, and how it moves after the handler

Current state

TriggerMake or n8n
StepsHappy path only
FailRetry or stop
PersonFinds out late, or never

Each automation is its own island. A failure is either retried blindly or lost.

Future state

TriggerUnchanged
StepsSame happy path
HandlerClassify, retry or park
Queue and alertOwner, payload, replay

The original steps stay in place. The handler can be removed from one automation without touching the others.

Future-state modules

Make / n8nSource automationsError handlerClassify and retryDestinationJob, invoice, CRMDead-letterAirtable queueAlertsSlack and emailHeartbeatMissed-run watch

How it is built

Module-level build
  1. Module AInventory

    Every automation is listed with owner, trigger, destination, last failure and what a silent failure costs.

  2. Module BClassification

    Each error is classed: rate limit, auth, validation, timeout, or unknown.

  3. Module CRetry policy

    Only rate limits and timeouts retry, on a written backoff, with a cap.

  4. Module DDead-letter queue

    Anything else is parked in Airtable with the full payload, the error and an owner.

  5. Module EIdempotent replay

    A successful replay cannot create a second job or invoice from the same payload key.

  6. Module FAlerts and heartbeats

    Critical failures alert at once. A missed heartbeat is treated as a failure of its own.

Field mapping

What the handler reads and writes
SourceField or signalDestinationRule
Make / n8n runExecution id, scenario or workflow name, timestampAirtable: Run keyBecomes the idempotency key
Error objectHTTP status, message, node or module nameAirtable: Error type and detailMapped to the classification table
Original inputFull JSON payloadAirtable: PayloadStored as received, never trimmed
ClassificationRetryable or notHandler pathRetry schedule, or park with no retry
Queue rowStatus, owner, attempt countSlack or emailAlert text and link; grouped in a 15 minute window
Heartbeat watchLast successful finishSlack and emailAlert if the gap exceeds the written maximum

Worked exhibits

Retry rules and one week of handled failures
Error typeRetry?ScheduleIf it still fails
Rate limit (429)Yes5 attempts, 1, 2, 4, 8, 16 minParked, High alert
Timeout / networkYes3 attempts, 2, 4, 8 minParked, High alert
AuthenticationNoPaused at onceParked, Critical alert
Validation / bad recordNoNo retryParked, named field, High alert
UnknownNoNo retryParked, Critical alert
Runs in the week12,48014 automations
Failed first try1861.5% of runs
Succeeded on retry141rate limits and timeouts
Parked in the queue45owner assigned
Duplicates created0idempotent replay
Rate limit92Timeout49Validation28Authentication11Unknown6
Dead-letter queue and the alert board
Airtable · dead-letter queue
ItemDL-1847
AutomationInvoice push, n8n
ErrorXero 429, rate limit
Attempts5 of 5
StatusParked
OwnerOperations
PayloadKept in full, 12 KB
Slack · operations channel
SeverityHigh
TextInvoice push parked 6 items. Top cause: Xero 429.
GroupingOne message per 15 min window
LinkOpens the Airtable view
Owner@ops-lead
DigestWeekly counts by automation

Exceptions and runbook

Runbook: who does what when something is parked
SignalThresholdGoes toFirst action
Critical failureLast retry failed on a money or customer flowFinance email and operations SlackOpen the item, fix the cause, replay once
Queue sizeMore than 10 open, or any older than 24 hoursOperations SlackReview the same morning
Heartbeat missedGap above the written maximumOwner and operations SlackCheck the trigger and the scheduling switch
Credential expiry14 days before, and on the dayOwner emailRenew and test the connection
Handler itself failsAnyOwner and a nominated second mailboxRestore the handler, then replay parked items

How to read the exhibits

  1. The findings are from the review of this kind of setup, not a claim about a named business.
  2. The rules table is the contract: rate limits retry, bad records do not.
  3. The week of figures shows the handler at work: most first-try failures recover, and nothing is created twice.

Test plan

PASS / FAIL checks before the next automation is wrapped
IDCheckMethodPASS whenFAIL when
T1Rate limitForce a 429 in a copyFive attempts on the schedule, then success or parkFewer attempts, or no park
T2AuthenticationRevoke a test tokenNo retry, paused, Critical alert at onceRetries, or no alert
T3ValidationSend a bad recordParked with the field namedRetried or lost
T4IdempotencyReplay a successful itemNo second record in the destinationA duplicate is created
T5HeartbeatSwitch off a test automationAlert inside the maximum gapNo alert
T6Alert groupingFail 20 items at onceOne grouped message per window20 separate messages
T7Replay after a fixFix the cause and replayItem created once, queue row closedItem stuck or duplicated
T8Handler failureBreak the handler in a copySecond-level email sentSilence

Scope

AreaIncluded
DiscoveryRead-only review of every Make scenario and n8n workflow, owners and failure history, written findings
InventoryEvery automation with owner, trigger, steps and failure cost
Shared error handlerOne pattern for Make and one for n8n, with types and retries
Dead-letter queueAirtable table, views, owner and status, with replay steps
Alerts and heartbeatsSlack and email by severity, grouped; missed-run watches
Test planThe PASS and FAIL checks in this write-up, run in a copy before go-live
Handover and supportWritten SOPs, a recorded walkthrough, files you own, 14 days of fixes, and an optional monthly hour block
Not includedRebuilding the automations themselves, a 24-hour on-call roster, or vendor subscriptions

Rollout, handover and monthly support

MilestoneDeliverableTimingShare of fee
M1 Discovery and rulesInventory, findings and error rules in writingWeek 125%
M2 Handler and queueHandler and queue built and tested in a copyWeek 230%
M3 RolloutHandler live on all automations, heartbeats onWeek 330%
M4 HandoverAlert tuning, runbook, SOPs, 14 days of fixesWeek 415%
Handover and monthly support

Handover

  • Written operating procedures for the happy path and each exception in the runbook
  • A recorded walkthrough of the architecture, the queue and how to replay an item
  • A credentials and ownership list, with every connection in your name
  • Fourteen days of fixes after go-live, for defects in the agreed scope

Monthly support

A set block of hours each month, at US$35 per hour, agreed in writing before it starts. Typical use of the block: watching the queue, renewing credentials, and adjusting retry windows after a vendor change

  • Review of the exception queue and the weekly digest
  • Small mapping or rule changes, agreed before they are made
  • A short written note on what was done and what is still open

Typical scope for this kind of project: US$1,500-3,000

About 4 weeks · fixed milestones. The systems health check (US$150, fixed) comes first and confirms the scope. Extra hourly work is US$35 per hour, agreed before it starts. After handover, a set block of hours each month is available at the same rate, agreed in writing, to watch the flows, handle changes and work the exception queue.

Discuss an engagement.

Outline the operational problem through the contact form. I respond with a recommended starting point, including whether the US$150 systems health check is the appropriate first step.

Related work snippets

Next step

Bring me the work that repeats.

I scope it in writing, test it against your current process on your own data and hand over everything I build.

Book a consultation