Pushka collects your errors, logs, traces, uptime checks, sessions and metrics, joins them on one trace id, and wires the result to alerts, on-call and a public status page.
Native SDKs for .NET · Node.js · the browser — one wire contract, no runtime dependencies, PII scrubbing on by default. Already on Sentry? Change one URL.
An error tracker, a log tool, an uptime checker and a metrics dashboard — each with its own bill, its own login and its own idea of what a request is. Pushka is one console that agrees with itself.
The SDKs stamp a single trace id on everything they send. The browser error, the Node span that caused it and the .NET log line underneath arrive already joined — no more matching timestamps across four tabs.
Pushka speaks the Sentry protocol. If your app already runs a Sentry SDK, point its DSN at Pushka and your errors start arriving — no re-instrumentation, no migration project.
250,000 signals a month on one shared allowance — errors, logs, spans, metrics, sessions and checks all draw from the same pool. No card, no expiry. Go over and Pushka sheds the spike instead of billing you.
The order things actually happen in — and which part of the console does the work.
The SDK sends the exception with its stack, the breadcrumbs before it, the release it shipped in and the user it happened to. Minified frames are symbolicated against the source map you uploaded on deploy — you get a file and a line, not a bundle offset.
Rules watch error volume, custom metric thresholds, uptime checks and the health of your outbound dependencies. Deduplication is claimed atomically, so two console replicas can never page you twice for the same thing.
The rotation knows who is awake. The escalation policy knows what happens if they aren't: step two after five minutes, a different channel, a different person. Acknowledging stops the chain.
The trace shows where the three seconds went. The replay shows the click that started it. The map shows which countries it reached. The release view shows the issue first appeared in 1.9.3.
The status page moved when the monitor did, and subscribers were notified. The error budget was debited against your SLA. The incident timeline is your postmortem draft.
Pushka ships an MCP server, so Claude Code, Cursor, Codex or Windsurf can query your issues, stacks, logs and traces while it writes the fix. It reads through a scoped API key — the agent can look at production; it cannot change it.
list_projectswhich projects this key can readsearch_issueswhat is broken, using the console's search syntaxget_issuethe exception, the symbolicated stack, the breadcrumbs, the trace idget_logsthe lines around it, already filtered to the traceclaude mcp add pushka \
-e PUSHKA_API_KEY=pk_live_… \
-- npx -y @pushkadotdev/mcpInstall the SDK, paste your project DSN, deploy. Errors, logs and request spans start arriving in the console already correlated — and PII scrubbing is on before you configure anything.
$ npm i @pushkadotdev/browserimport Pushka from "@pushkadotdev/browser";
Pushka.init({
dsn: "https://api.pushka.dev/api/v1/envelope?project_key=pk_live_…",
environment: "production",
release: "web@1.9.4",
});
Pushka.setUser({ id: "u_88213" });
Pushka.logger.info("checkout started", { cartId: "c_99", items: 3 });Turn off the ones you don’t use — for the whole organization, or for one person.
Retention you can read off the plan, and paid plans start with a 14-day trial — no card.
Everything a solo developer needs to watch one project. No card, no expiry.
Start freeIncludes logs, traces, the service map, sessions, releases, custom metrics, source maps and the audit trail.
For a team shipping to production: on-call, escalation, SLA and agent access.
Start 14-day trialAdds SLA tracking, scheduled reports, shared links and 2 check regions.
Higher volume, longer retention, the change log, AI observability and SMS/voice.
Start 14-day trialAdds AI/LLM observability, the customer-facing change log, SMS and voice alerts, and a status page on your own domain.
Negotiated volume and retention, SSO/SCIM, white-label status pages.
Talk to usAdds SAML SSO, SCIM provisioning, audit export, white-label status pages and priority support.
A signal is one error, log line, span, metric point, session or uptime check. They share a single allowance, so there aren’t six separate meters to forecast. Go over it and Pushka sheds the spike and tells you — rather than billing you for it.
A companion app for iOS and Android: incidents, monitors, issues and on-call — with push notifications that deep-link straight to the problem, and acknowledge in one tap.
Instrument something this afternoon and you’ll know about the next incident before your customers do. The free plan doesn’t expire.