Logs, metrics, traces — wired up by default.
Every request through tGate emits a structured log line, a span, and contributes to your latency histograms. Query in the dashboard or stream out to Datadog, Honeycomb, or your S3 bucket.
See your p99 move in real time.
Every PoP samples 100% of requests, builds quantiles in-memory, and pushes deltas to the dashboard every 250ms. Slice by route, status code, region, IdP group, or any policy attribute.
- ✓True quantilesDDSketch-based, not approximated from buckets.
- ✓By any dimensionGroup by tag with no pre-aggregation step.
- ✓Compare to last weekBuilt-in delta view spots regressions instantly.
Every request carries a trace, end to end.
tGate emits OpenTelemetry-compatible spans for each hop — TLS, policy eval, edge cache, origin call. Connect your origin to the same trace via W3C traceparent.
- ✓OTel-nativeSend to any collector. Or query in tGate directly.
- ✓Tail-based samplingKeep all errored or slow traces; downsample the rest.
- ✓Linked logsClick a span → see every log line that fired during it.
Stop instrumenting auth boilerplate, gateway hops, retries.
The gateway is the perfect place to capture every request. Don't reinvent it in every service.
import { trace } from '@opentelemetry/api';
import { Resource } from '@opentelemetry/resources';
import { NodeSDK } from '@opentelemetry/sdk-node';
// ...30 imports
const sdk = new NodeSDK({
resource: new Resource({ 'service.name': 'api' }),
traceExporter: new OTLPTraceExporter({ url }),
metricReader: new PeriodicExportingMetricReader({ ... }),
instrumentations: [...autoInstrumentations()],
});
sdk.start();
// per service. Per language. Per environment.# every tunnel emits OTel automatically.
# add custom tags from policy:
route "/*" {
trace.tag("tenant", req.user.tenant)
trace.tag("plan", req.user.plan)
}Inspect any request, replay any failure.
Every request is loggable, traceable, and replayable from the dashboard CLI.
Live tail
Stream filtered logs to your terminal in real time.
tgate logs tail --grep 5xxReplay requests
Re-run any logged request against a different env.
tgate replay rq_8x2...Saved queries
Pin dashboards. Share via URL with team filters preserved.
tgate query save ...Alerts
Threshold + anomaly alerts to PagerDuty, Slack, webhook.
alert if p99 > 200msExport to S3
Stream raw events to your bucket in Parquet.
destination = s3://...Synthetic checks
Continuous probes from every PoP, alerting on regression.
tgate synth add ...We turned off our $48k/yr APM contract three weeks after switching to tGate. The traces are better — they include the edge — and there's no per-host pricing trap.
Stop paying for what your gateway already knows.
Logs, metrics, traces — included on every plan, kept for a year, queryable in milliseconds.