Error log triage

Help me find the cause of this error.

Error and log lines:
{{log}}

What changed recently, and where it happens:
{{context}}

1. Say in one sentence what the log shows for certain, separately from what it only suggests.
2. List up to three likely causes, most likely first. For each, give the evidence in the log or context that points to it, and what would rule it out.
3. For each cause, give the quickest check that confirms or rules it out: a command, a query or a place to look.
4. A safe way to limit the damage while I check, if there is one (for example rolling back or switching off the new feature).
5. The fix for the most likely cause, marked as depending on its check.
6. What extra information would change your ranking.

Do not present a guess as the root cause.

How to use it

Paste the error with a few lines before it, plus what changed recently (deploys, config, traffic) and where it happens (one user, one region, everywhere). Remove secrets and personal data first. Run the checks in the order given.

Example

log

2026-10-20T14:02:11Z app[web.2]: GET /api/reports 200 1843ms 2026-10-20T14:02:13Z router: at=error code=H12 desc="Request timeout" method=GET path="/api/reports?range=12m" dyno=web.1 service=30001ms status=503 2026-10-20T14:02:13Z app[web.1]: [pg] Connection terminated unexpectedly 2026-10-20T14:02:14Z app[web.1]: Error: Connection terminated unexpectedly at Connection.<anonymous> (/app/node_modules/pg/lib/client.js:132:73) 2026-10-20T14:02:20Z app[web.2]: GET /api/reports?range=12m 200 9120ms 2026-10-20T14:03:02Z router: at=error code=H12 desc="Request timeout" method=GET path="/api/reports?range=12m" dyno=web.2 service=30000ms status=503

context

Customers report occasional 502 errors on the Reports page since Monday's deploy, which added a 12-month range option. Other pages are fine. Hosted on Heroku with a managed Postgres add-on.