[Remote Config] Real-time listener stops receiving updates, with no onError, when the stream starts during a persisted backoff

Open Beginner friendly
#8,827 1 comment 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
2/5
Estimated time
1-3 hours
Newbie friendliness
78/100
Issue type
Bug
Clarity
Clearly specified
Activity status
Active
Tech stack
android, java
Domain
backend

Research direction

Read ConfigRealtimeHttpClient.beginRealtimeHttpStream() and the retry path it calls when a persisted backoff is active. Confirm how the connection-running flag affects canMakeHttpStreamConnection(), then check the existing Remote Config real-time stream tests for coverage of that path. Done means the stream retries after backoff ends and the listener can receive updates.

Written by the indexing model from the issue text.

Description

api: remoteconfig
[READ] Step 1: Are you in the right place?

Yes, this is a bug in firebase-config in this repository.

[REQUIRED] Step 2: Describe your environment
  • Android Studio version: **Android Studio Quail 4 | 2026.1.4
  • Firebase Component: Remote Config (real-time updates, addOnConfigUpdateListener)
  • Component version: 23.1.0
[REQUIRED] Step 3: Describe the problem

The app stops receiving Remote Config updates for as long as the process lives. Once this happens, a registered ConfigUpdateListener never fires again: values changed in the console never reach the app, and onError is never called, so the app can't detect or recover from it. Only killing the process brings updates back.

On Android a process can stay alive for days, so apps that rely on real-time updates (rather than frequent fetch() calls) keep running on stale config for that whole time. In our app, users stayed on old values for days.

When it happens: two conditions, in this order.

  1. A stream connection fails while the app is in the foreground. This means no network, or the server answering 408, 429, 502, 503 or 504. The SDK then persists a backoff end time in SharedPreferences. The backoff is about 2 minutes after the first failure (randomized ±50%) and doubles with each consecutive failure, up to 256 minutes. It survives process restarts and is cleared only by the next successful (200) connection.
  2. Before that backoff ends, the SDK starts the stream again. It does this whenever the app calls addOnConfigUpdateListener (for example on a cold start) or comes back to the foreground. Both go through startHttpConnection() → beginRealtimeHttpStream(), which hits the backoff early return described below and leaves the stream stuck.

Both are routine. For example, the user opens the app on a flaky connection, switches away, and comes back a minute later. From then on the app receives no updates until the process is killed.

Steps to reproduce:
  1. Register a ConfigUpdateListener with the app in the foreground.
  2. Make the real-time stream connection fail while the app is in the foreground. Keep the app in the foreground until the attempt fails.
  3. Restore network access, kill the app, and relaunch it while the backoff is still in effect. Register the listener again.
  4. After the backoff end time has passed, publish a Remote Config change.

Expected: when the backoff ends, the stream reconnects and the listener receives the update.
Actual: no connection is attempted after the backoff ends, the listener receives nothing, and onError is never called.

Relevant Code:

In ConfigRealtimeHttpClient.beginRealtimeHttpStream(), the method sets isHttpConnectionRunning up front. The early return for an active backoff never clears it, because only the finally block of the connection task clears it:

public void beginRealtimeHttpStream() {
  if (!checkAndSetHttpConnectionFlagIfNotRunning()) { // sets isHttpConnectionRunning = true
    return;
  }

  ConfigSharedPrefsClient.RealtimeBackoffMetadata backoffMetadata =
      sharedPrefsClient.getRealtimeBackoffMetadata();
  Date currentTime = new Date(clock.currentTimeMillis());
  if (currentTime.before(backoffMetadata.getBackoffEndTime())) {
    // No connection was opened, but isHttpConnectionRunning is still true here.
    // The retry below goes through canMakeHttpStreamConnection(), which requires
    // !isHttpConnectionRunning, so it is silently dropped and the stream stays stopped.
    retryHttpConnectionWhenBackoffEnds();
    return;
  }
  // ...
}

Suggested fix: release the flag before scheduling the retry, since no connection was opened on this path:

setIsHttpConnectionRunning(false);
retryHttpConnectionWhenBackoffEnds();
return;
Dominant language
Java
Stars
2.6k
Forks
733
Avg merge
2d 15h
Merged PRs (30d)
72

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from firebase/firebase-android-sdk

All issues in firebase/firebase-android-sdk

Similar issues

More Java issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.