Obics – Automated Observability Cost Reduction

TL;DR

TL;DR: Obics is an observability cost-reduction agent that connects to your Datadog/Grafana/Splunk instance and GitHub repo, identifies telemetry bloat in your source code, estimates the savings, and opens a PR to fix it.

Source and Accuracy Notes

⚠️ This section is MANDATORY. All links must be verified from actual source, not guessed.

• Project page: obics.io
• Launch HN thread: news.ycombinator.com/item?id=49087681
• License: Proprietary (free tier available; source code not publicly licensed)

What Is Obics?

Obics is an automated observability cost-reduction tool. It connects to your existing observability platform (Datadog, Grafana, Splunk) and your source control repository, then daily analyzes:

• Telemetry volume and cost groupings
• Logs back to their source code location
• Dashboards, alerts, and other consumers that depend on that telemetry
• Optimization opportunities with estimated savings

When a fixable issue is found, Obics opens a pull request with the source-code change.

The core problem Obics addresses: telemetry debt. Logs, metrics, and traces accumulate for years. Debug logs ship to production. Incident-specific logging outlives the incident. Auto-instrumentation and third-party libraries emit spans everyone ignores. High-cardinality metric tags quietly multiply bills. Nobody cleans it up because safely removing telemetry requires knowing whether anything depends on it — context that usually isn’t worth the trouble.

How It Works

Step 1: Connect Your Platforms

Obics integrates with your observability vendor and your GitHub repository. The setup authenticates against your Datadog, Grafana, or Splunk account and grants read access to your source code.

Step 2: Daily Analysis

Every day, Obics runs a set of automated checks:

1. Volume and cost analysis — groups telemetry by cost impact
2. Similarity mapping — clusters duplicate or near-duplicate logs
3. Source-code mapping — traces telemetry back to the exact file and line
4. Dependency detection — checks dashboards, alerts, and other consumers
5. Opportunity identification — surfaces actionable optimization candidates

Step 3: PR or Platform Action

For source-code fixable issues, Obics opens a pull request directly in your repo. For third-party library issues, retention policy problems, or log-generated metrics, it instead configures exclusion filters or modification rules using your observability platform’s native capabilities.

What It Detects

Based on the launch post, Obics identifies these issue types:

• Duplicate logs — identical or near-identical logging statements
• Forgotten debug logging — INFO/DEBUG level logs that should have been removed before shipping
• Unnecessary metric labels — high-cardinality tags multiplying time series counts
• Equivalent metrics — multiple metrics measuring the same thing
• Repetitive logging — loops or hot paths generating excessive log volume
• PII exposure — logs containing personally identifiable information
• Excessive object serialization — costly serialization with no observability value

How Obics Differs from Existing Approaches

| Approach | What It Does | Limitation | |---|---|---| | Sampling | Discards telemetry to reduce ingest | Trades visibility for lower bills | | Tiered storage | Moves old data to cheaper storage | Slower queries, rehydration complexity | | Telemetry pipelines (Cribl, Edge Delta) | Filters telemetry at the collector | Applications still generate the same volume; CPU/egress waste remains | | Switching vendors | Moves to a cheaper platform | Expensive migration; non-equivalent features | | Raising log level | Blunt WARN-only filter | Poor proxy for usefulness; valuable INFO logs discarded | | Obics | Fixes the source code generating waste | Handles third-party cases via platform config |

The key distinction: most existing approaches treat the symptom (cost), not the root cause (code generating unnecessary telemetry). Obics modifies the source, and when that is not possible, falls back to platform-native configuration.

Practical Evaluation Checklist

• [ ] Sign up at obics.io and connect a trial Datadog/Grafana/Splunk account
• [ ] Authorize the GitHub integration for your main application repo
• [ ] Review the first analysis run — check whether detected issues are genuine
• [ ] Verify PR descriptions accurately identify the source file and line
• [ ] Test the estimated savings figure against your actual observability bill
• [ ] Merge a low-risk PR (debug log removal) and confirm the change behaves as expected
• [ ] Check that dashboards and alerts are not broken after PR is merged

Security Notes

• Obics requires read access to your observability platform and source code
• It opens PRs rather than pushing directly to main — every change is reviewed before merge
• For PII detection: the feature scans log content, consider whether that aligns with your data handling policy

FAQ

Q: Which observability platforms does Obics support? A: Datadog, Grafana, and Splunk at launch. Support for additional platforms may be added based on demand.

Q: Does Obics push changes directly, or does it always open a PR? A: For source-code issues, Obics opens a pull request. For platform-level configuration (exclusion filters, metric label removal), it applies the change directly using the observability vendor’s API.

Q: Can it break my dashboards or alerts? A: Obics detects dashboard and alert dependencies before recommending removal. It prefers source-code fixes that preserve the telemetry signal while reducing volume.

Q: Is there a free tier? A: Yes, Obics offers a free tier with limited analysis scope. Pricing details are on the project website.

Conclusion

Obics tackles observability cost reduction differently — by going to the source rather than filtering at the border. For teams running significant Datadog, Grafana, or Splunk bills, it offers a zero-effort audit with concrete PRs to act on. The daily analysis loop means it catches new telemetry debt as it accumulates, not just at initial setup.

If you manage a production system with growing observability costs, it is worth connecting a trial account and reviewing the first analysis report.

• Project page: obics.io
• Launch thread: news.ycombinator.com/item?id=49087681