SDK burst guard protects projects from duplicate storms. Instead of letting one noisy source create thousands of nearly identical events, the SDK can enter cooldown, keep only every Nth duplicate, and emit guard telemetry so you know suppression happened.
How Burst Guard Works
- The SDK counts duplicate logs or errors over a rolling window.
- Once the configured threshold is crossed, the runtime enters cooldown.
- During cooldown, only every Nth duplicate is kept.
- A guard signal is emitted so Errova can record that suppression happened.
JavaScript Configuration Example
import { init } from "@errova/sdk"init({dsn: process.env.ERROVA_DSN ?? "",dsnSecret: process.env.ERROVA_DSN_SECRET,burstGuardDefaults: {windowSeconds: 60,cooldownSeconds: 600,logs: {enabled: true,threshold: 100,sampleEveryDuringCooldown: 200,},errors: {enabled: false,threshold: 20,sampleEveryDuringCooldown: 50,},},burstGuardOverrides: {errors: {enabled: true,threshold: 10,sampleEveryDuringCooldown: 5,},},guardSourceIdentifier: "api:billing-worker",})
Python Configuration Example
from errova_sdk import initinit(dsn="https://ingest.example.com/ingest/<public_key>/events",dsn_secret="<one-time-secret-from-key-create>",guard_source_identifier="worker:billing",burst_guard_overrides={"logs": {"threshold": 1,"sample_every_during_cooldown": 999,}},)
Base Defaults in the JavaScript SDK Family
- Window: 60 seconds
- Cooldown: 600 seconds
- Logs: enabled, threshold 100, sample every 200 duplicates during cooldown
- Errors: disabled by default, threshold 20, sample every 50 duplicates during cooldown
Operational Guidance
- Fix the noisy code path at the source instead of relying on suppression forever.
- Use a stable
guardSourceIdentifierper service or worker so telemetry stays meaningful. - Enable alert rules for burst-guard triggers when duplicate storms should page the team.
- Most users do not need to call
/guard-signalsmanually because SDKs already do it.