Two parts. Work the decision honestly, then plan the monitoring you would need before committing anything.
Before you commit stake
- Write down your capital and compare it against the protocol minimum. State plainly whether you can also accept delegation, and how you intend to attract any.
- Price archive access for every chain you intend to index, monthly, as a real number.
- Compute what the same capital would earn delegated to an existing indexer, doing nothing. That is the number you must beat.
- Pick three subgraphs you would allocate to and write one sentence each on why, without referring to their signal rank.
- Open a dashboard and check whether those three are already crowded with allocated stake.
- List the four silent failure modes from this collection and say, for each, what alert you would build.
- Decide your max_amount_willing_to_lose_grt and write down the reasoning, not just the number.
- Name the person who covers your alerts when you are unavailable. If there is no name, that is your finding.
An honest closing note
Where next
- The procedures this collection deliberately omits: official indexing documentation.
- Live indexing status, POI comparison and performance data: Lodestar.
- Who sets the parameters you have been reading, and how they change: governance.
- What the people signalling on your subgraphs are thinking: curator.