Monitor a Solana program on five fronts: failed transactions, authority and upgrade changes, balances of your keeper and fee-payer wallets, oracle freshness, and how far your RPC lags behind the chain. Send each alert to a named person with a written runbook, and test the alerts before mainnet, not after the first incident.
| Signal | Where it comes from | Alert when |
|---|---|---|
| Failed transactions | Transaction logs and signature statuses | The failure rate of one instruction jumps, or a new error code appears |
| Upgrade and authority changes | The program data account and authority accounts | Any change you did not schedule |
| Keeper and fee-payer balances | Account balances of your own wallets | Balance can no longer cover the next days of fees |
| Oracle freshness | Timestamps in the price feed accounts you read | The feed is older than your own limit |
| RPC lag and errors | Slot compared with a second source, error and rate-limit responses | Your endpoint falls behind or starts rejecting calls |
Set the thresholds from your own traffic. A number that suits another protocol may be wrong for yours.
- Launch day is when monitoring matters most. Build it before mainnet, not after the first incident.
- Watch changes you did not make: an upgrade, a new authority or a large transfer out of a vault.
- Your own wallets can fail the product. A keeper that runs out of SOL stops liquidations silently.
- Read data at confirmed or finalized commitment, and compare your RPC with a second source.
- An alert nobody owns is noise. Give each one a named person and a short runbook.
Most Solana teams spend months on the program and an afternoon on monitoring. Then a transaction starts failing at three in the morning, or a keeper wallet runs dry, and the first person to notice is a user on social media. Monitoring is part of the launch, and it is cheaper to set up while the team still remembers how the program works.
1. Failed transactions
Failed transactions still land onchain and still pay fees, so they leave a trail. Track the failure rate per instruction and the error codes behind it. A single error is normal. A sudden rise for one instruction, or a new error code after a deploy, deserves a person. Anchor programs publish custom errors in the IDL, which makes the codes readable.
2. Upgrade and authority changes
A program that can be upgraded stores its upgrade authority in the program data account. Watch that account and every authority your program uses, such as admin keys and mint or freeze authorities. Any change you did not schedule is an incident until proven otherwise. Pair the alert with a multisig on the upgrade authority, so one stolen key cannot ship new code.
3. Balances you own
- Keeper and bot wallets: liquidators, crankers and reward distributors pay fees from their own SOL. When it runs out, the product stops without an error on your side.
- Fee payer for user transactions: if you sponsor fees, watch the balance and the daily burn.
- Vaults and treasuries: alert on large or unusual transfers out, and on balances that move outside the pattern you expect.
4. Oracle freshness
If your program reads a price feed, check how old the last update is. A stale price can let a bad trade or liquidation through. Decide your own limit, enforce it in the program, and alert when the feed gets close to it.
5. Your RPC and data pipeline
- Slot lag: compare the slot your provider reports with a second source. A node that falls behind returns old state without an error.
- Errors and rate limits: count rejected calls and throttling responses, and see who hits the limit first.
- Commitment level: read at confirmed or finalized for anything that moves money. Processed data can be dropped.
- Indexer lag: if you run an indexer, alert when it falls behind the chain head, so users do not see stale numbers.
How to get the data
Solana RPC supports WebSocket subscriptions for accounts, program accounts and logs, which is enough for a small program. Heavier setups stream account and transaction updates through Geyser-based gRPC feeds, or read from an indexer. See Solana indexer vs RPC for when to add one, and how to choose an RPC provider for what to ask about uptime and failover.
Mistakes to avoid
- No owner: an alert that goes to a channel nobody reads is the same as no alert.
- Alerting on everything: too many alerts teach the team to ignore them. Start with the five above.
- Untested alerts: trigger each one on devnet before launch.
- One data source: a single RPC endpoint can be wrong or down at the moment you need it.
- No runbook: write down the first three steps for each alert while nobody is under pressure.
Monitoring is one line in our Solana pre-launch security checklist. You can browse verified providers on the Solana infrastructure providers page.
Frequently asked questions
What should I monitor on a Solana program?
Failed transactions per instruction, upgrade and authority changes, balances of your keeper and fee-payer wallets, oracle freshness, and how far your RPC lags behind the chain. Start with these five and add more only when you need them.
How do I get alerts on a Solana program?
Small programs can use RPC WebSocket subscriptions for accounts and logs. Larger setups stream updates through Geyser-based gRPC feeds or read from an indexer, then send alerts to the person who owns each one.
How do I know if someone upgraded my program?
Watch the program data account, which stores the upgrade authority, and the authority accounts your program uses. Alert on any change you did not schedule, and put the upgrade authority behind a multisig.
Should I test my monitoring before mainnet?
Yes. Cause each failure on devnet on purpose and confirm that the alert fires, reaches a named person and comes with a short runbook.
Find a Solana infrastructure provider
Describe your workload and stage. Smart Match suggests verified infrastructure and development providers with a reason for each match. Free for teams.
Get matched →Looking for help with this? Browse Solana infrastructure and RPC providers, or describe your project and get matched for free.
Related reading
- How to choose a Solana RPC provider: 10 questions to askHow to pick a Solana RPC provider: shared vs dedicated nodes, rate limits, streaming, transaction landing, failover and security, with questions to ask and a way to test.
- Solana indexer vs RPC: what to use, and when to add an indexerWhen RPC is enough on Solana and when you need an indexer: history, filters and totals, Geyser streaming, decoding, build or buy, and questions to ask a provider.
- Solana program upgrade authority: how to secure it with a multisigWho should hold a Solana program upgrade authority, how to move it to a multisig, when to add a timelock or make the program immutable, and the mistakes to avoid.