kubectl-multi-logs

Tail logs from multiple Kubernetes pods simultaneously — real-time or historical — with a Homebrew-style live progress display.

Features

Parallel execution: Pod lookups, container fetches, and log streams run concurrently via goroutines.
Replica coverage: Automatically includes all replicas of each app.
Real-time & historical modes: Follow logs live or retrieve from a relative time (10m, 1h, 2d).
Previous-instance logs: Pass -p to also collect logs from the container instance that ran before the last restart — useful for crash diagnostics.
Flexible filtering: Case-insensitive pattern matching with | OR support; dedicated -e flag for errors.
Organised output: Lines prefixed with [pod:container] (or [pod:container:prev] for previous instances).
Accurate time ranges: Each kubectl logs call uses --timestamps; the real log timestamp is parsed and used for the HH:MM:SS → HH:MM:SS summary range. The timestamp prefix is stripped before writing to the log file so output stays clean.
Brew-style live progress: Scroll-and-advance terminal UI — each item prints a permanent ✔ line as it completes, grouped by app, with a live spinner for remaining streams.
Fixed semantic coloring: Pod names are always cyan; container names are plain — consistent and easy to scan without distracting per-app palette shifts.
Exit summary with timestamps: On completion a per-container summary table shows pod, container, line count, and HH:MM:SS → HH:MM:SS first/last log timestamps. Previous streams are labelled :prev; pods that have never restarted show – no previous instance.
Debug command log: CLI output (start banner through summary) is always mirrored to command.log in the same directory as the binary. The file is overwritten on each run.
Single binary: No shell runtime required — just kubectl on $PATH.

Installation

Prerequisites: Go 1.21+ and kubectl configured on $PATH.

git clone https://github.com/vianhanif/kubectl-multi-logs.git
cd kubectl-multi-logs
go build -o kubectl-multi-logs .

To install as a kubectl plugin (enables kubectl multi-logs):

cp kubectl-multi-logs /usr/local/bin/kubectl-multi-logs

Shell alias

mylogs () {
    ~/path/to/kubectl-multi-logs "$@" app-1 app-2 app-3
}

Usage

kubectl-multi-logs [-n namespace] [-s since] [-T timeout] [-g pattern] [-e] [-p] [-o [output_file]] <app1> <app2> ...

Options

Flag	Description	Default
`-n`	Kubernetes namespace	current context
`-s`	Historical logs since (e.g. `10m`, `1h`, `2d`)	follow live
`-T`	Per-stream timeout in collect mode (`-s`); `0` = no limit	`2m`
`-g`	Filter lines — case-insensitive, `\|` for OR	none
`-e`	Filter for `ERROR\|WARN\|Exception\|failed\|error`	off
`-p`	Also collect logs from the previous container instance (before last restart)	off
`-o`	Output file name (`-o` alone uses default)	`tail_multiple_logs_data.log`
`-verbose`	Show per-pod/container rows during progress (default: compact 3-bar view)	off

Examples

To discover available app names in your cluster:

kubectl get pods -o jsonpath='{.items[*].metadata.labels.app}' | tr ' ' '\n' | sort | uniq

Follow live logs from agent-service and tez-api:
```
kubectl-multi-logs agent-service tez-api
```

Last 30 minutes of logs, errors only:

kubectl-multi-logs -s 30m -e agent-service

Last 30 minutes, including logs from before the last restart:
```
kubectl-multi-logs -s 30m -p agent-service
```
Crash investigation — errors from current and previous instance:
```
kubectl-multi-logs -s 1h -e -p agent-service tez-api
```

Filter errors in production namespace:

kubectl-multi-logs -n production -e tez-api

Search for a pattern across apps and save to custom file:

kubectl-multi-logs -g ERROR -o custom_logs.log agent-service tez-api

Use -o alone to write to the default file:

kubectl-multi-logs -s 15m -o agent-service tez-api

Example Output

Terminal Output

The tool runs through three phases with live-updating progress bars (go-pretty StyleBlocks, green █ fill). By default all three phases are shown as compact single-line bars. Pass -verbose for per-item detail.

Default — compact 3-bar view (one bar per phase, all live at once):

Mid-run (phase 2 in progress):

Collecting from last 10m  ·  saving to /path/to/tail_multiple_logs_data.log
  Finding pods... (3/3)                    ...  [  1.342s]
  Fetching containers... (16/21)           ...  ║██████████████████▒░░░░░░░░░░░░░░░░░║  [  2.1s]
  Collecting logs...                       ...  ║░░░░░░░░░░░░░░░░▒█▒░░░░░░░░░░░░░░░░║  [  2.1s]

Completed:

Collecting from last 10m  ·  saving to /path/to/tail_multiple_logs_data.log
  Finding pods... (3/3)                    ...  [  1.342s]
  Fetching containers... (21/21)           ...  [  2.801s]
  Collecting logs... (23/23)               ...  [ 10.450s]
✔ All streams finished  total elapsed: 15s

Bars are green (█); they refresh live then become static lines once each phase finishes. The indeterminate spinner (▒█▒) is shown while the total is not yet known.

With -verbose — per-item rows appear as results stream in:

Phase 1 — completed (overall bar replaced by static line, items below it):

  Finding pods for apps...  3 / 3           ...  [  1.342s]
  agent-service                                            15 pod(s)  ...  [  312ms]
  agent-service-worker                                      3 pod(s)  ...  [  456ms]
  tez-api                                                   3 pod(s)  ...  [  623ms]

Phase 2 — completed:

  Fetching containers for pods...  21 / 21  ...  [  2.801s]
  agent-service-deployment-7b54bf66c7-2f79w      1 container(s)  ...  [  92ms]
  agent-service-deployment-7b54bf66c7-425mh      1 container(s)  ...  [ 108ms]
  agent-service-worker-deployment-6b5b6cbdf7-bbkll  2 container(s)  ...  [ 134ms]
  tez-api-deployment-77d5fb6445-hknxx            1 container(s)  ...  [ 187ms]
  ...

Phase 3 — collecting logs, grouped by app:

  Collecting logs… (21 / 21 done)          ...  [ 10.450s]
  agent-service
    [agent-service-deployment-7b54bf66c7-2f79w] agent-service   collected · 1557 lines  ...  [  9.2s]
    [agent-service-deployment-7b54bf66c7-425mh] agent-service   collected · 1551 lines  ...  [  9.8s]
  agent-service-worker
    [agent-service-worker-deployment-6b5b6cbdf7-bbkll] agent-cron-worker   collected · 153 lines  ...  [  8.1s]
    [agent-service-worker-deployment-6b5b6cbdf7-bbkll] agent-service-worker  collected · 61 lines  ...  [  7.5s]
  tez-api
    [tez-api-deployment-77d5fb6445-hknxx] tez-api               collected · 14504 lines  ...  [ 10.2s]

Pod names are displayed in cyan; container names are in plain text — fixed colours regardless of which apps are being tailed.

Summary (printed on completion or Ctrl+C):

── Summary

╭──────────────────────────────────────────────────────────────────┬────┬───────┬──────────────────────╮
│                                                                  │    │ Lines │ Time range           │
├──────────────────────────────────────────────────────────────────┼────┼───────┼──────────────────────┤
│ agent-service  2 pod(s) · 2 stream(s) · 3108 lines              │    │       │                      │
│   agent-service-deployment-7b54bf66c7-2f79w                      │    │       │                      │
│     agent-service                                                │ ✔  │  1557 │ 10:12:03 → 10:22:41  │
│   agent-service-deployment-7b54bf66c7-425mh                      │    │       │                      │
│     agent-service                                                │ ✔  │  1551 │ 10:12:04 → 10:22:39  │
├──────────────────────────────────────────────────────────────────┼────┼───────┼──────────────────────┤
│ agent-service-worker  1 pod(s) · 2 stream(s) · 214 lines        │    │       │                      │
│   agent-service-worker-deployment-6b5b6cbdf7-bbkll               │    │       │                      │
│     agent-cron-worker                                            │ ✔  │   153 │ 10:12:05 → 10:22:10  │
│     agent-service-worker                                         │ ✔  │    61 │ 10:12:05 → 10:20:44  │
├──────────────────────────────────────────────────────────────────┼────┼───────┼──────────────────────┤
│ tez-api  1 pod(s) · 1 stream(s) · 14504 lines                   │    │       │                      │
│   tez-api-deployment-77d5fb6445-hknxx                            │    │       │                      │
│     tez-api                                                      │ ✔  │ 14504 │ 10:12:01 → 10:22:58  │
├──────────────────────────────────────────────────────────────────┼────┼───────┼──────────────────────┤
│ TOTAL  3 app(s) · 4 pod(s) · 5 stream(s)                        │    │ 17826 │                      │
╰──────────────────────────────────────────────────────────────────┴────┴───────┴──────────────────────╯

The hierarchy is app → pod → container. Failed streams show ✗ with a truncated error message instead of line counts and timestamps.

When -p is active each container gets an additional row labelled container:prev. If the pod has never been restarted that row shows – no previous instance instead of line counts:

│ my-app  1 pod(s) · 2 stream(s) · 846 lines                      │    │       │                      │
│   my-app-deployment-7b54bf66c7-abc12                             │    │       │                      │
│     my-app:prev                                                  │ -  │     0 │ no previous instance │
│     my-app                                                       │ ✔  │   846 │ 10:12:03 → 10:22:41  │

If the pod has been restarted the previous instance shows its own line count and time range:

│     my-app:prev                                                  │ ✔  │   312 │ 09:58:14 → 10:09:57  │
│     my-app                                                       │ ✔  │   846 │ 10:12:03 → 10:22:41  │

Log File Output

Each log line is prefixed with [pod:container]. When -p is used, previous-instance lines carry a :prev suffix so they are trivially distinguishable when grepping:

[agent-service-5d4b8f6c9-abc12:agent-service:prev] 2026-02-18 10:08:41 ERROR OOMKilled — container exceeded memory limit
[agent-service-5d4b8f6c9-abc12:agent-service] 2026-02-18 10:23:01 INFO  Starting request processing for policy #98712
[agent-service-5d4b8f6c9-abc12:agent-service] 2026-02-18 10:23:01 INFO  Fetching agent data for agent_id=4521
[agent-service-5d4b8f6c9-xyz34:agent-service] 2026-02-18 10:23:02 INFO  Starting request processing for policy #98713
[tez-api-6f8b9c7d4-xk2mn:tez-api] 2026-02-18 10:23:02 INFO  Received webhook callback for policy #98712
[tez-api-6f8b9c7d4-xk2mn:celery-worker] 2026-02-18 10:23:03 INFO  Processing task: send_notification policy_id=98712
[agent-service-5d4b8f6c9-abc12:agent-service] 2026-02-18 10:23:03 ERROR Failed to fetch provider response: timeout after 30s
[tez-api-6f8b9c7d4-pq5rs:tez-api] 2026-02-18 10:23:04 WARN  Retrying request to agent-service (attempt 2/3)
[agent-service-5d4b8f6c9-xyz34:nginx] 2026-02-18 10:23:04 10.0.0.1 - - "POST /api/v1/policies HTTP/1.1" 504 0
[tez-api-6f8b9c7d4-xk2mn:celery-worker] 2026-02-18 10:23:05 ERROR Task failed: send_notification ConnectionError
[tez-api-6f8b9c7d4-pq5rs:tez-api] 2026-02-18 10:23:05 INFO  Successfully retried request for policy #98712

The RFC3339 timestamp injected by kubectl --timestamps is parsed for the summary time range and then stripped before writing, so the log file always contains clean application log lines.

Tip: This output file can be shared with GitHub Copilot or other AI tools to analyze and correlate errors across pods.

Bash version

The original bash implementation is kept at versions/bash/tail_multiple_logs.sh for reference. It has the same flags and behaviour but uses background subshells and temp files for concurrency.

Code layout

File	Responsibility
`config.go`	All tunable constants — timeouts, truncation widths, timestamp format, progress update frequency
`main.go`	Flag parsing, output-file setup, signal handler, top-level orchestration
`kubectl.go`	Domain types (`appPod`, `podContainers`) and `kubectl` helper functions
`phases.go`	Parallel pod/container discovery workers (`findPods`, `fetchContainers`) and the four phase runners (`runPhase1`, `runPhase2`, `runPhase1Clean`, `runPhase2Clean`)
`stream.go`	`streamState`, `streamConfig`, `streamLogs`, `matchesPattern`, `launchStreams`
`progress.go`	`newPW`, `phaseMonitor`, `displayMonitor`, `displayMonitorClean`, `cleanLabel`, `activeStop`
`summary.go`	`printSummary` — rounded table grouped by app → pod → container using go-pretty/table

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Navigation Menu

Repository files navigation

kubectl-multi-logs

Features

Installation

Shell alias

Usage

Options

Examples

Example Output

Terminal Output

Log File Output

Bash version

Code layout

Contributing

License

About

Releases

Packages

Contributors

Languages

Name		Name	Last commit message	Last commit date
.serena		.serena
versions/bash		versions/bash
.gitignore		.gitignore
README.md		README.md
config.go		config.go
go.mod		go.mod
go.sum		go.sum
kubectl.go		kubectl.go
main.go		main.go
phases.go		phases.go
progress.go		progress.go
stream.go		stream.go
summary.go		summary.go

Latest commit

History

Folders and files

Repository files navigation

kubectl-multi-logs

Features

Installation

Shell alias

Usage

Options

Examples

Example Output

Terminal Output

Log File Output

Bash version

Code layout

Contributing

License

About

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages