Solana ships Microscope, an open-source monitoring tool for on-chain programs

Solana has released Microscope, a self-hosted tool that turns a program's on-chain transactions into dashboards and alerts, according to a developer post published on Solana.com dated October 5, 2026. The code is on GitHub under the MIT license.

The pitch is narrow and operational. Microscope reads confirmed transactions from a program — via a Yellowstone gRPC stream or a Solana RPC endpoint — and uses the program's IDL to decode its instructions and events. It writes structured records to Loki and metrics to Prometheus, which Grafana then renders as dashboards and evaluates against alert rules. Alerts can go to Slack, Telegram, or PagerDuty. One deployment watches one program and, optionally, one Squads multisig that controls it, including proposal approvals.

Because the data is decoded rather than matched against raw log text, rules can target a specific instruction or event and inspect a field such as a transferred amount. The dashboard surfaces transaction and instruction error rates, decoded instructions by type, recent events, and multisig activity. Per the announcement, a failed on-chain instruction shows up as program activity and does not indicate the indexer failed.

Microscope also monitors itself. It ships health alerts for its own pipeline — covering a stalled stream, stale RPC polls, a corrupt checkpoint, or records that decode but never reach Loki — so an empty dashboard can be diagnosed as "program is quiet" versus "data stopped flowing." Every alert carries a signal_kind label distinguishing on-chain activity from a datasource problem.

Deployment options listed include a Docker Compose stack, AWS or GCP VMs via Terraform, a Grafana Cloud connection, and Kubernetes reference manifests. The repository includes a "setup-deployment skill" meant to help an AI agent prepare and validate the configuration, plus a Jupiter Lend walkthrough with sample alerts.

Key facts

• Microscope is open-source, self-hosted, MIT-licensed; code on GitHub (Solana.com developer post, Oct 5, 2026).
• Decodes program instructions and events via IDL; data sources are Yellowstone gRPC or Solana RPC (same post).
• Stack: Loki for records, Prometheus for metrics, Grafana for dashboards and alerting; outputs to Slack, Telegram, PagerDuty (same post).
• One deployment monitors one program and optionally one Squads multisig (same post).
• Deployable via Docker Compose, Terraform on AWS/GCP, Grafana Cloud, or Kubernetes (same post).

The real-world read

This is a product announcement on Solana's own channel, so the framing is Solana's — read the capability claims as the vendor's description, not independently tested results. That said, the substance is checkable: it's MIT-licensed code on GitHub, and the design leans on standard, auditable components (Grafana, Loki, Prometheus) rather than a proprietary black box. No pricing, adoption numbers, or benchmarks are given, and none are claimed. The Jupiter Lend example is described as a demonstration walkthrough, not a customer endorsement. There is no second, independent source on this release yet, so performance at scale is unverified.

This is informational reporting, not financial or technical advice.

Sources

• Solana.com — "Introducing Solana Microscope" (developer post, Oct 5, 2026): primary and only source; provided all product details, architecture, and deployment options. This is the vendor's own announcement — treated as a product launch, not independent reporting.