README
¶
kubectl logz
The aim of this tool is to provide a middle-ground between a Enterprise log facility (like Splunk) and tailing
container logs using the clunky and limited kubectl logs command.
Solve these problem:
- Find problems.
- Give context, not just the error.
- Do it fast.
Log Structure
There are hundreds of different log formats. The parser must normalize log messages into a semantic format.
| field | description |
|---|---|
time |
RFC339 timestamp. |
level |
error, warn, info, or debug. |
msg |
Human readable text. |
threadId |
The thread or Coroutine ID. |
This information can be contextualized, to find the exact thread that executed the code
| field | description |
|---|---|
host |
The process host. |
pid |
The PID. |
(host,pid,threadId) gives a GUID for work being done.
Requests may traverse processes, recording OpenTelemetry fields:
| field | description |
|---|---|
traceId |
Trace ID. |
spanId |
Span ID. |
It and does so by trying each of the following parsers:
- Structured:
- logfmt, e.g.
time=2022-12-04T18:36:52Z level=warn msg="thing happened" - JSON, e.g.
{"time":"2022-12-04T18:36:52Z","level":"warn","msg"="thing happened"}
- logfmt, e.g.
- Unstructured:
- Space-separated fields, e.g.
2022-12-04T18:36:52Z [warn] thing happened". - Failover.
- Space-separated fields, e.g.
Addendum
HTTP access logs include the following:
| field | description |
|---|---|
ip |
IP address of the client. |
user |
Client user name. |
method |
HTTP method. |
path |
HTTP path. |
status |
HTTP status code. |
Documentation
¶
There is no documentation for this package.
Source Files
Directories
Click to show internal directories.
Click to hide internal directories.