Skip to main content

Sysdig Secure

The Sysdig connector pulls runtime / cloud-detection events from the Sysdig Secure Events API — every Falco-style policy hit raised by the same engine that powers the open-source falco connector, plus the higher-level posture detections Sysdig layers on top.

If you already run the open-source falco connector for self-hosted clusters, the Sysdig connector is what you add for cross-account / multi-cluster Sysdig SaaS visibility without standing up a per-cluster Falco sidecar.

What you get

Event familyExamples
Workload policy hitsRead sensitive file untrusted, Disallowed binary in container, Network connection to unfamiliar destination
Cloud policy hitsAWS / GCP / Azure posture rules (e.g. public S3 bucket created)
Container driftImage substitution, runtime image hash diverges from registry
Identity policyService-account token mounted outside expected namespace
CNAPP correlationsSysdig's own multi-signal CDR (cloud detection-and-response) entries

Events are normalized with source: sysdig, category: siem (runtime detection plane).

Prerequisites

  • A Sysdig Secure subscription (Sysdig Secure Plus or Enterprise — the free Sysdig Monitor tier does not expose Secure Events).
  • A Sysdig API token scoped to read access on Secure Events. Tokens are minted under Settings → API → Tokens.
  • Know your Sysdig region prefix (visible in the URL of any Sysdig page):
Region prefixHostname
us1secure.sysdig.com
us2us2.app.sysdig.com
us3app.us3.sysdig.com
us4app.us4.sysdig.com
eu1eu1.app.sysdig.com
au1app.au1.sysdig.com
me2app.me2.sysdig.com

Setup walkthrough

1. Mint the API token

  1. Sysdig Secure → Settings → API → Tokens → New Token.
  2. Name it aisoc-readonly so it's auditable.
  3. Scope: Secure Events: Read only. Do not grant Posture: Write or Compliance: Write — they are unnecessary for ingestion.
  4. Copy the token immediately. Sysdig will not show it again.

2. Add the connector in AiSOC

  1. Connectors → Add connector → Sysdig Secure.
  2. region = the prefix from the table above.
  3. api_token = the token from step 1 (stored encrypted in the credential vault).
  4. Test connection → calls GET /api/v1/secureEvents?limit=1.
  5. Save.

Polling details

  • Default interval: 300 seconds (configurable down to 60s and up to 3600s).
  • The connector queries GET /api/v1/secureEvents with a from/to window in nanoseconds covering the last since_seconds window plus a 30-second overlap to absorb out-of-order event delivery.
  • Pagination walks both offset/limit and cursor styles depending on which the tenant's API version honours.
  • An upper bound of 25 pages × 100 events = 2,500 events per poll prevents the connector from monopolising an API quota during a burst — additional events are picked up on the next poll cycle.

Severity mapping

Sysdig event severities use the Falco syslog ladder (0..7, lower = more severe). The connector folds this onto the AiSOC scale:

Sysdig severityAiSOC severity
0 (Emergency) / 1 (Alert)critical
2 (Critical) / 3 (Error)high
4 (Warning)medium
5 (Notice)low
6 (Informational) / 7 (Debug)info (suppressed by default — flip the env var AISOC_SYSDIG_KEEP_INFO=1 to keep them)

Troubleshooting

HTTP 401: invalid token — token was revoked / rotated, or you copied the Monitor API token rather than the Secure API token (they share the UI but the scopes differ).

HTTP 403: insufficient permissions — token lacks the Secure Events: Read scope. Re-mint with the right scope; do not escalate the existing token.

Connection timeout — region prefix wrong. The connector resolves the prefix to a fixed hostname (see table above); a US1 token will not authenticate against the EU1 endpoint.

No events returned despite Sysdig dashboard showing detections — the API only surfaces events from policies tagged for Secure Events output. Posture findings on un-tagged policies stay inside the Sysdig UI.

  • Falco — self-hosted runtime detection. Use this if you run open-source Falco and want the raw policy hits without the Sysdig SaaS subscription.
  • Kubernetes Audit — control-plane audit log. Pairs well with Sysdig for cluster-level forensics (control + data plane in one investigation).