Skip to main content
Symptom-first, most common first. Every behaviour here is deliberate — the fix is usually one property or one concept, not a bug report. If your symptom isn’t listed, ask in Discussions.

​
The app exits immediately at startup

The log ends with:
Alginte fails fast by design when the broker is unreachable — a clear startup error instead of a UI where every page times out. Almost always one of:
  • Wrong address for the container’s point of view. Inside Docker, localhost:9092 is the container, not your machine. Use a hostname reachable from the container, or join the broker’s Docker network — see Run with Docker.
  • The broker isn’t up yet. In compose files, order the app after the broker with a healthcheck (depends_on: condition: service_healthy) — the playground compose shows the pattern.
  • No broker at all? The playground bundles one.
The escape hatch (alginte.startup-connectivity-check.enabled=false) skips the check entirely; the app starts regardless and connects once the broker appears — use it when the broker is expected to come up later.

​
An area is missing from the menu

Kafka Connect, Schema Registry, or ksqlDB don’t appear in the sidebar. Their menu areas show only when the matching URL is configured (ALGINTE_CONNECT_URL, ALGINTE_SCHEMAREGISTRY_URL, ALGINTE_KSQLDB_URL) — an unconfigured integration is hidden rather than shown broken. Wire them up in Configuration. If the URL is set but the service is down, the area stays visible and degrades; the service-health indicator in the sidebar footer shows which upstream is unreachable.

​
Messages: fewer records than requested

Two different causes, distinguished by the note above the table:
  • “Incomplete data polled” (warning) — the polling budget ran out before the fetch filled. Raise alginte.consumeMaxPollAttempts (default 40) for slow brokers or very large records. The read was genuinely cut short.
  • Compacted topic (info, from 0.7.0) — nothing was cut short: on a compacted topic offset ranges contain gaps, so the range simply held fewer live records than its size suggests. Blank value cells there are tombstones (real records marking a key deleted), not decoding failures.

​
A stream shows ERROR (or won’t leave REBALANCING)

From 0.7.0: a stream whose processing threads have died shows ERROR — most commonly a source topic deleted after deploy, which Kafka Streams cannot recover from. Recreate the topic (or fix the design) and Stop/Start the stream — details in Deploy & manage. Deploying against a topic that doesn’t exist is blocked at Submit with a message naming it. On 0.6.0 and earlier, that same dead-thread condition displays as REBALANCING indefinitely — if a stream rebalances forever, check the server log for MissingSourceTopicException.

​
OIDC sign-in is rejected with license_required

Single-sign-on is a PRO feature; a Community instance rejects OIDC sign-in with license_required. Community’s built-in login is the single-account option. Full matrix on Editions & licensing; OIDC specifics (including the allowlist fail-fast at startup) on OIDC setup.

​
Port 8888 is already taken

Another process (Jupyter is the classic) holds the port. For the playground: ALGINTE_PORT=8890 docker compose up -d. For your own docker run, change the host side of the mapping: -p 9000:8888.

​
Still stuck?

Open a Discussion or an Issue — include the Alginte version (About page), broker distribution/version, and the relevant log lines.