Skip to content
RocketFlag Docs

FAQ & Troubleshooting

My flag returns 404 Not Found. Why?

Section titled “My flag returns 404 Not Found. Why?”

A 404 from the evaluation API means one of:

  • The Flag ID is wrong — copy it again from the flags table in the console.
  • The flag is a Group (Multi-Environment) Flag and you didn’t pass env, or the env value doesn’t match a defined environment label. The env parameter is case-sensitive. See Group Flags.

I renamed an environment and now I get 404.

Section titled “I renamed an environment and now I get 404.”

Environment labels are matched exactly. If you rename a label in the console, update the env parameter in your application code to match. See Environment Management.

My cohort request returns 400 Bad Request.

Section titled “My cohort request returns 400 Bad Request.”

The cohort value contained characters that weren’t URL-encoded (a common culprit is +). Always URL-encode cohort values before sending them. The SDKs handle this for you. See the API Reference.

The same user sometimes sees the feature and sometimes doesn’t.

Section titled “The same user sometimes sees the feature and sometimes doesn’t.”

Percentage rollouts are sticky, but only when RocketFlag has a key to bucket on. Send a stable identifier as targetingKey (or cohort) on every evaluation request and a user keeps the same result across requests and across environments. If you send neither, every request is an independent random roll, which is the usual cause of a flickering feature. The SDKs send targetingKey when you include it in the context. See Sticky rollouts.

Also check that the identifier really is stable: a value that changes per session, such as an anonymous session ID that resets, gives a different bucket each time. If you need certain people always on, add them to Always on for (Cohorts).

My audience doesn’t match.

Section titled “My audience doesn’t match.”

Audience matching is exact and case-sensitive, and an attribute that is not sent never matches, not even is not one of. Use the Try it panel in the audience editor to paste your query string and see which rule matched. Also confirm the attribute key is spelled the same in your code (keys are case-sensitive) and that it is not one of the reserved keys cohort, env or targetingKey. See Audiences.

My analytics numbers look behind.

Section titled “My analytics numbers look behind.”

Analytics are processed in batches. Usage counts update frequently, but there can be a delay of up to 10–15 minutes before the latest hits appear in the charts. Analytics require the Teams tier or above. See Analytics & Stats.

Can I cache flag responses?

Section titled “Can I cache flag responses?”

Yes. The Node, React, and Go SDKs support opt-in in-memory caching with a configurable TTL, keyed by flag ID and user context. Caching is off by default.

I don’t see the Caretaker tab / analytics / audit logs / SAML SSO.

Section titled “I don’t see the Caretaker tab / analytics / audit logs / SAML SSO.”

These are tier-gated:

  • Caretaker (stale-flag detection) and Analytics require the Teams tier or above, on a project that belongs to an Organisation.
  • Audiences require the Teams tier or above. An organisation below Teams can still see its existing audiences on the Audiences tab, read-only, but cannot create, edit or delete them. See Audiences.
  • Organisation-wide Audit Logs and SAML SSO require the Enterprise tier.

See Organisation Tiers.

Which SDK should I use?

Section titled “Which SDK should I use?”
  • React apps: the React SDK (provider + useFlag hook + <Flag> component).
  • Node.js / other JavaScript: the Node SDK.
  • Go services: the Go SDK.
  • Anything else: call the API directly — it’s a single GET request.

My question isn’t here.

Section titled “My question isn’t here.”

Reach out through the RocketFlag website — we’d love to help.