Google's Data Manager diagnostics guidance recommends waiting before the first status check, increasing intervals with a 1.3 backoff, adding jitter, capping intervals at 60 minutes, and stopping after a maximum 24-hour period. A worker that ignores those boundaries can create rate pressure or leave conversions in an ambiguous state forever.
Use this for WordPress lead uploads, WooCommerce conversions, CRM-qualified leads, agency connectors, scheduled batches, dashboards, and support alerts using Data Manager API request status.
Quick answer
Persist the request ID and send evidence before polling. Wait at least the documented initial processing period, schedule each next check from durable state with 1.3 backoff and jitter, cap the interval at 60 minutes, and stop normal polling at 24 hours. Move the request to `stale_needs_owner`, preserve destination status and reason counts, and avoid resending events until deduplication and business impact are understood.
What to check first
- Store request ID, destination, event count, immutable event IDs, send time, code version, consent state, redacted batch reference, first poll, last poll, and current status.
- Calculate next poll from durable timestamps so restarts and duplicate workers cannot reset the backoff or schedule the same request twice.
- Test the initial wait, 1.3 multiplier, random jitter, 60-minute cap, 24-hour stop, 429 Retry-After, transient network errors, and terminal success, partial success, or failure.
- Separate API processing state from business state so an operator can see whether conversions were sent, acknowledged, reported, retried, or still ambiguous.
- At 24 hours, alert one owner with request evidence, destination details, safe reason counts, deduplication policy, support path, and a clear decision deadline.
Diagnostic table
Use this table to keep the work practical. It connects the symptom to evidence and a verification step.
| Action | Evidence to collect | How to verify |
|---|
| Persist the request lifecycle | Store request ID, destination, event count, immutable event IDs, send time, code version, consent state, redacted batch reference, first poll, last poll, and current status. | Restarts do not reset elapsed time or create parallel polling. |
| Implement durable backoff and jitter | Calculate next poll from durable timestamps so restarts and duplicate workers cannot reset the backoff or schedule the same request twice. | Intervals grow with jitter and never exceed the configured 60-minute cap. |
| Cap concurrent polling | Test the initial wait, 1.3 multiplier, random jitter, 60-minute cap, 24-hour stop, 429 Retry-After, transient network errors, and terminal success, partial success, or failure. | Success, partial success, and failure stop future polls and trigger the correct workflow. |
| Stop at the 24-hour boundary | Separate API processing state from business state so an operator can see whether conversions were sent, acknowledged, reported, retried, or still ambiguous. | The 24-hour alert contains actionable evidence and no raw personal data. |
Why this usually happens
- Stateless workers can forget previous intervals after a restart.
- Several schedulers can poll the same request concurrently.
- HTTP retry behavior can be confused with asynchronous processing state.
- A stale request may be resent before destination deduplication is proven.
Useful command or data shape
Adapt paths, IDs, and privacy handling to the site before running commands or storing data on production.
request_id: req-9f81
sent_at: 2026-07-28T10:00:00Z
status: PROCESSING
poll_attempt: 11
next_interval_minutes: 60
jitter_seconds: 83
elapsed_hours: 24.1
worker_state: stale_needs_owner
automatic_resend: false
Safe fix order
Do the work in a sequence that makes each result easy to prove. Stop if a step produces new evidence that changes the incident scope.
- Persist the request lifecycle
- Implement durable backoff and jitter
- Cap concurrent polling
- Stop at the 24-hour boundary
- Escalate without automatic replay
Decision rule
Pass when each request has one durable schedule, polling follows documented bounds, terminal states stop immediately, and a 24-hour nonterminal request becomes an owned decision without duplicate event effects.
What to tell the client or owner
Give the owner the affected versions, exact workflow, observed result, business impact, evidence location, temporary control, named owner, and next review time. Remove credentials and personal data from shared screenshots and logs.
Production verification checklist
- Restarts do not reset elapsed time or create parallel polling.
- Intervals grow with jitter and never exceed the configured 60-minute cap.
- Success, partial success, and failure stop future polls and trigger the correct workflow.
- The 24-hour alert contains actionable evidence and no raw personal data.
Mistakes to avoid
- Do not change several plugins, cache rules, or infrastructure settings before preserving a baseline.
- Do not treat one successful test as proof for retries, alternate clients, background work, or mixed-version fleets.
- Do not paste secrets, personal data, complete production payloads, or customer files into tickets or screenshots.
- Do not close the test until the final user-visible state and the server-side evidence agree.
Questions teams ask during testing
Should the events be resent at 24 hours?
Not automatically. First determine destination state and whether stable event IDs guarantee deduplication without changing reporting or bidding.
Can `Retry-After` replace the status schedule?
Honor applicable HTTP guidance, but keep transport retries and asynchronous request-status polling as separate state machines.
When HandL WP should help
Bring in HandL WP when a production checkout, form, editor, security gate, media pipeline, or paid lead workflow is at risk. We can preserve evidence, isolate the failing layer, make the smallest corrective change, and verify the result across WordPress, connected services, logs, and the user journey.
If this is active on a production site, operate Data Manager requests safely.
Related HandL WP guides
Use these related guides when the same issue touches tracking, security, checkout, or crawler visibility.
Helpful references