Issues

Quickly understand what requires your attention and drive your investigations

The issues page is a useful place to start a troubleshooting or investigation flow from. It gathers together all active issues found in your Kubernetes environment.

Issue Types

• HTTP / gRPC Failures Capturing failed HTTP calls with Response Status Codes of:

  • 5XX — Internal Server Error

  • 429 — Too Many Requests

• MySQL / PostgreSQL Failures

  Capturing failed SQL statement executions with Response Errors Codes such as:

  • 1146 — No Such Table

  • 1040 — Too Many Connections

  • 1064 — Syntax Error

• Redis Failures Capturing any reported Error by the Redis serialization protocol (RESP), such as:

  • ERR unknown command

• Container Restarts Capturing all container restart events across the cluster, with Exit Codes such as:

  • 0 — Completed

  • 137 — OOMKilled

• Deployment Failures

  Capturing events such as:

  • MinimumReplicasUnavailable — Deployment does not have minimum availabiltiy

Issue Aggregation

Issues are auto-detected and aggregated - representing many identical repeating incidents. Aggregation help cutting through the noise quickly and reach insights like when a new type of issue started to appear, and when it was last seen.

Issues are grouped by:

• Type (HTTP, gRPC, Container Restart, etc..)

• Status Code / Error Code (e.g HTTP 500, gRPC 13)

• Workload name

• Namespace

Issues aggregated by Type, Status Code, Workload and Namespace

The smart aggregation mechanism will also identify query parameters, remove them, and group the stripped queries / API URIs into patterns. This allows users to easily identify and isolate the root cause of a problem.

The smart aggregation of the API /product/{product_id}

Troubleshooting with Issues

Each issue is assigned a velocity graph showing it's behavior over time (like when it was first seen) and a live counter of its number of incidents.

By clicking on an issue, users can access the specific traces captured around the relevant issue. Each trace is related to the exact resource that was used (e.g. raw API URI, or SQL query), it's latency and Status Code / Error Code.

Traces captured around the relevant issue

Further clicking on a selected captured trace allows the user to investigate the root cause of the issue with the entire payload (body and headers) of the request and response, the information around the participating container, the application logs around incident's time and the full context of the metrics around the incident.

Investigation with the context of logs and metrics around the incident