DEV Community

Cover image for Monitoring Lioran S3 with Health Checks, Metrics and the CLI
Swaraj Puppalwar
Swaraj Puppalwar

Posted on

Monitoring Lioran S3 with Health Checks, Metrics and the CLI

#observability #devops #monitoring #storage

Monitoring Lioran S3

A storage service that only tells you “running” or “dead” is not especially cooperative.

Lioran S3 V1 Pre-Alpha exposes health, system diagnostics, performance information, and Prometheus-style metrics.

The project is developed by Lioran Developer Solutions under Lioran Group, led by Swaraj Puppalwar.

Health endpoint

Basic liveness:

curl   https://storage.example.com/health
Enter fullscreen mode Exit fullscreen mode

A load balancer or orchestrator can use this as a simple probe.

Driver health 

const health =
  await client.health();

console.log(health);
Enter fullscreen mode Exit fullscreen mode

Use it during startup diagnostics or deployment verification.

System information 

const info =
  await client.systemInfo();

console.log(info);
Enter fullscreen mode Exit fullscreen mode

The system information API is useful for inspecting runtime/storage configuration without SSHing into the box for every question.

Performance API 

const perf =
  await client.systemPerformance();

console.log(perf);
Enter fullscreen mode Exit fullscreen mode

This can expose storage-side write latency and throughput information.

Metrics 

const metrics =
  await client.metrics();

console.log(metrics);
Enter fullscreen mode Exit fullscreen mode

The metrics endpoint uses Prometheus text exposition.

Typical categories include:

  • request counts
  • HTTP latency
  • throughput
  • active uploads
  • object counts

CLI health 

liorans3 system health
Enter fullscreen mode Exit fullscreen mode

CLI system info 

liorans3 system info
Enter fullscreen mode Exit fullscreen mode

CLI performance 

liorans3 system performance
Enter fullscreen mode Exit fullscreen mode

Alias:

liorans3 system perf
Enter fullscreen mode Exit fullscreen mode

CLI metrics

Raw:

liorans3 system metrics
Enter fullscreen mode Exit fullscreen mode

JSON-friendly mode where supported:

liorans3 system metrics   --json
Enter fullscreen mode Exit fullscreen mode

Doctor

The broader diagnostic command is:

liorans3 doctor
Enter fullscreen mode Exit fullscreen mode

doctor is intended to help inspect:

  • local configuration
  • selected profile
  • credential availability
  • connectivity
  • server reachability
  • common setup failures

Credentials should remain masked in diagnostic output.

A practical monitoring stack

A small deployment can look like:

Lioran S3
   |
   +--> /health
   |
   +--> /metrics
           |
           v
      Prometheus
           |
           v
        Grafana
Enter fullscreen mode Exit fullscreen mode

Your application monitoring should also track client-side symptoms.

Server metrics alone will not tell you that:

  • DNS is failing from one region
  • a reverse proxy changed timeout behavior
  • clients are retrying too aggressively
  • users are saturating a single network path

Alerts worth having

At minimum:

  • node unreachable
  • health probe failing
  • low disk space
  • rapid storage growth
  • elevated write latency
  • elevated error rate
  • authentication failure spike
  • multipart failure spike
  • unusual bandwidth saturation

Disk alerts

Storage infrastructure deserves two different disk alarms:

Headroom

Warn before the server's own low-watermark rejects writes.

Growth rate

Alert if free space is disappearing much faster than normal.

Capacity planning is easier when you know both current fullness and slope.

Benchmarking

The TypeScript SDK repository also includes benchmark tooling for write, read, and mixed workloads.

Benchmarks are useful for regression detection, but do not turn one throughput number into a universal product claim.

Record:

  • CPU
  • RAM
  • disk type
  • filesystem
  • durability mode
  • concurrency
  • object sizes
  • TLS/reverse proxy path
  • network topology
  • software version

Without the test environment, benchmark numbers become decorative numerology.

JSON output for automation

CLI automation should prefer:

liorans3 ... --json
Enter fullscreen mode Exit fullscreen mode

Then:

liorans3 bucket ls   --json   | jq '.'
Enter fullscreen mode Exit fullscreen mode

Machine-readable output lets CI reason about data instead of terminal formatting.

Pre-alpha operations

V1 Pre-Alpha is specifically the phase where observability is valuable because behavior is still being exercised under increasingly realistic workloads.

Measure first.

Optimize second.

Guessing is cheaper only until it reaches production.

Docs: https://docs.liorans3.sbs

Top comments (0)