Pulse vs OpenSearch Doctor: Which Is Right for Your Team?

If you're looking for a monitoring or observability tool for your OpenSearch cluster, you've likely come across both Pulse (by Bonsai) and OpenSearch Doctor. They sound like they solve the same problem. They don't — or rather, they solve it from different angles for different audiences.

This post is an honest comparison. We're the team behind OpenSearch Doctor, so take that for what it's worth — but we'll try to give you a fair picture of what each product actually does, what it costs, and who should use which.

What Pulse Is

Pulse is Bonsai's monitoring product for Elasticsearch and OpenSearch clusters. It's designed primarily for teams who want a polished UI to view cluster metrics: node stats, index sizes, query latency, shard distribution, and so on. It connects to your cluster and presents the data in clean dashboards.

Pulse is a good product. The UI is well-designed, setup is reasonably straightforward, and it covers the core metric categories you'd want to see. If you want a Grafana-style dashboard without building one yourself, Pulse delivers that.

What OpenSearch Doctor Is

OpenSearch Doctor is a diagnostic tool, not a metrics dashboard. The distinction matters.

A metrics dashboard shows you numbers: heap is at 87%, disk is at 73%, search latency averaged 450ms over the last hour. You then have to interpret those numbers and decide what they mean. Is 87% heap alarming or normal for this cluster? Is 450ms latency a problem or expected given the query load? The dashboard doesn't tell you.

OpenSearch Doctor runs structured diagnostic checks and surfaces findings: specific problems detected in your cluster, explained in plain language, with recommended remediation steps. Not "heap is 87%" but "Node es-data-01 heap is at 87% — GC pressure is likely causing latency spikes. The old generation GC is running every 8 seconds. Consider force-merging historical indices or adding a data node."

It's the difference between a thermometer and a doctor.

Side-by-Side Comparison

Architecture

Pulse: SaaS with cloud-side data collection. Pulse's infrastructure connects to your cluster endpoint from outside. Your cluster must be reachable over the internet (or via a network tunnel), and you provide credentials that Pulse stores and uses to scrape metrics.

OpenSearch Doctor: Agent-based. A lightweight Go binary runs on your server, connects to OpenSearch locally (or via localhost), and pushes diagnostic results to the cloud dashboard. Your cluster never needs to be internet-accessible. The agent reads metrics and config data — it never reads document content. Source code is open for review.

What It Checks

Pulse: Core infrastructure metrics — heap, CPU, disk, JVM GC, search and indexing throughput, node availability. Broadly: the metrics you'd get from GET /_nodes/stats.

OpenSearch Doctor: Infrastructure metrics plus OpenSearch-specific diagnostics: ISM policy failures, snapshot health and recency, security configuration (anonymous access, TLS status), shard allocation root cause analysis, index template conflicts, read-only index detection, over-sharding, fielddata abuse, cluster settings anomalies, and more. Checks that require calling and interpreting multiple OpenSearch APIs, not just reading a single stats endpoint.

Alerting

Pulse: Alert rules based on metric thresholds. Email notifications. Some plans include Slack and PagerDuty.

OpenSearch Doctor: Finding-based alerts — you get notified when a specific problem is detected, not when a raw metric crosses a threshold. Email, Slack, and webhook support. Alert rules can be scoped per cluster or per finding type.

Pricing

Pulse: Starts at around $15–30/month per cluster depending on the plan, scaling up based on cluster size and features. No permanent free tier for production use.

OpenSearch Doctor: Free for 1 cluster (no credit card required), $39/month for up to 3 clusters (Starter), $99/month for up to 10 clusters (Pro), $199/month for unlimited clusters (Scale). Each tier includes alerting.

Cluster Accessibility Requirement

Pulse: Your cluster must be reachable from Pulse's scraper infrastructure. This typically means the cluster is internet-accessible or you maintain a VPN/tunnel.

OpenSearch Doctor: The agent runs where your cluster runs. Clusters can be fully private — on-premise, air-gapped VPC, localhost-only. No inbound network access required.

Who Should Use Pulse

• Teams who want a polished metrics dashboard and are comfortable exposing their cluster endpoint to an external service
• Teams where the cluster is already internet-accessible (hosted on a cloud provider with a public endpoint)
• Teams who need time-series graphs and want to slice metrics over time windows
• Teams who are already Bonsai customers and want monitoring integrated into that ecosystem

Who Should Use OpenSearch Doctor

• Teams running OpenSearch in a private VPC, on-premise, or on bare metal where the cluster is not internet-accessible
• Teams who want actionable findings, not just raw numbers — "here's what's wrong and how to fix it" rather than "here's the heap graph"
• Teams dealing with OpenSearch-specific issues: ISM policies, shard allocation problems, snapshot failures, security configuration — problems that require interpreting API responses, not just scraping stats
• Teams who want to review the monitoring code before running it on their infrastructure
• Teams managing multiple clusters at different price points

Can You Use Both?

Yes, and for some teams it makes sense. Pulse gives you time-series metric graphs that OpenSearch Doctor doesn't focus on. OpenSearch Doctor gives you diagnostic findings and OpenSearch-specific checks that Pulse doesn't cover. If you want both — a metrics dashboard and a diagnostic layer — they're genuinely complementary rather than redundant.

That said, for most self-managed OpenSearch teams, the biggest gap isn't metric graphs (you can get those free with Prometheus + Grafana) — it's the diagnostic layer that tells you why something is wrong and what to do about it. If you're choosing one, that's the more common unmet need.

The Honest Summary

If you want pretty metric dashboards and your cluster is already internet-accessible: Pulse is a solid choice.

If you want to know what's actually wrong with your cluster and what to do about it — especially for ISM, snapshots, security config, and shard allocation issues that raw metrics can't surface — OpenSearch Doctor is built specifically for that.

If you're not sure which you need: start with OpenSearch Doctor's free tier (1 cluster, no credit card). Run it for a week against your cluster. The diagnostic findings alone will tell you whether you have the kind of problems it's designed to catch.