Events not showing in AllStak: diagnose it in minutes
Your SDK is installed but the dashboard stays empty, or new errors stop arriving. Almost every case comes down to one of six checks below.
What this looks like
You initialized the AllStak SDK, deployed, triggered a test error — and nothing appears in the project. Or events used to flow and suddenly stopped after a deploy, a key rotation, or an infrastructure change. The SDK itself rarely logs anything visible when sends fail, so the dashboard staying empty is usually the only symptom.
Events also "disappear" when they actually arrived — but under a different environment than the one your dashboard is filtered to, or as handled events while you are only looking for crashes. Work through the steps in order; each one rules out a whole class of causes.
Common root causes
API key wrong or not loaded
The key in your code points at a different project, or the environment variable that should hold it is empty in the running process — common after deploys that don't carry the variable across.
Environment filter hides the events
Events land under the environment string the SDK sent (e.g. development), while the dashboard is filtered to production. The data is there — you are just not looking at it.
SDK init runs too late or not at all
If errors are thrown before the SDK initializes — or the init code path never executes in production builds — there is nothing in place to capture them.
Network egress or plan limits
A firewall, proxy, or TLS interception blocking HTTPS to the ingest endpoint silently drops events. Separately, sustained high volume can be throttled under your plan limits.
Step-by-step diagnosis
Run these checks in order — each rules out the most common cause before the next.
- 1
Verify the API key value
Open your project settings in the AllStak dashboard and compare the key character-for-character with what your app uses. Keys look like ask_live_… and are scoped per project — a key from another project sends your events elsewhere.
- 2
Confirm the env var is loaded at runtime
Log the presence (not the value) of the variable at startup — e.g. whether ALLSTAK_API_KEY is set. A variable defined in your local .env but missing from the deploy environment is the single most common cause of an empty dashboard.
- 3
Check the environment filter in the dashboard
Events appear under the environment your SDK sent (the environment option in init or config). Switch the dashboard's environment filter to the value your SDK is actually configured with — including a default like development if you never set it.
- 4
Make sure init runs before errors happen
Initialize the SDK as early as possible in your entry file — before importing the rest of the app where supported (e.g. AllStak.init at the top of a Node.js entry, instrumentation.ts in Next.js, allstak.init() before creating the Python app).
- 5
Test network egress to the ingest endpoint
From the machine running your app, confirm HTTPS requests to the AllStak ingest endpoint succeed — corporate proxies, container network policies, and TLS-intercepting middleboxes all silently drop SDK traffic.
- 6
Check capture gates and plan limits
Unhandled exceptions are captured automatically, but handled ones need an explicit captureException call — and some integrations have capture gates (e.g. allstak.capture-exceptions) that may have been disabled. If volume is very high, ingest can also be throttled under plan limits.
Prevent it from recurring
- Inject the API key from an environment variable in every deploy environment — never hardcode it.
- Set environment explicitly in SDK init so staging and production never mix or hide each other.
- Initialize the SDK first in your entry file, before any code that can throw.
- After every deploy, trigger one known test error and confirm it arrives before moving on.
- Keep an eye on plan usage so throttling never surprises you during an incident.
Still stuck?
If you have worked through all six steps and events still are not arriving, check the AllStak status page first to rule out a platform-side issue, then email [email protected] with your project name, SDK and version, and roughly when you expected events — we will trace it from the ingest side.
Frequently asked questions
How quickly should events appear after they are sent?
Events normally appear in near real-time — within seconds of the SDK sending them. If you are waiting minutes, something in this guide is blocking them; events do not queue up for delayed delivery on the dashboard side.
Why do my events show under the wrong environment?
The dashboard groups events by the environment string the SDK sent. If your production deploy still passes environment: 'development' (or never sets it), events land there. Fix the environment option in init and redeploy.
Are handled exceptions captured automatically?
No. Unhandled exceptions are automatic; handled ones inside try/catch need an explicit captureException call. If you only see crashes but not caught errors, that is by design — add manual captures where you need them.
Can plan limits cause events to be dropped?
Sustained volume beyond your plan limits can be throttled. If events arrive normally at low volume but go missing during spikes, review your plan usage in the dashboard.
Explore more
Compare
Get your events flowing
One API key, one init call, and your errors, logs, and traces land in one dashboard. Start free and see your first event in minutes.