patrick-ogrady / snowplow Goto Github PK

Write a script that prepares a new Ubuntu Google Cloud host to run snowplow.

We should use tar on unix instead of zip to preserve file attributes (source).

We should modify the backup command to store a SHA of the compressed staking key alongside the encrypted backup to ensure the staking key we retrieve in restore is correct.

We should install the Google Cloud monitoring Agent in ./scripts/setup.sh to get better visibility into memory and disk usage.

We should give avalanchego some threshold to recover before sending an unhealthy message. This should be configurable via .snowplow.yaml. Even a basic handler for unhealthy timeouts would be sufficient at this stage.

snowplow should support backing up the data volume in Google Cloud Storage so that new validators don't need to re-sync the entire chain from scratch.

Remove explicit call to get checksum from cmd. This should be done automatically in storage.

Periodically, we should send a status message with how many peers and the tips of each chain?

Write a script that prepares a new Ubuntu Google Cloud host to run snowplow.

We should add commands to backup a provided config (it may contain API tokens that are annoying to re-populate).

The health monitor should be built as discrete goroutines all uploading last "successful" checks to a struct. Then there should be a master loop that determines if a message should be sent.

In the case of a regression from healthy to not healthy, we should provide the option to call a phone number instead of just text messaging.

We should add the snowplow logo to the social preview for the repository.

Optionally consider the validator healthy once it has a minimum number of peers (via some configurable value in .snowplow.yaml).

We should add a command that prints out the NodeID of the staking credentials kept at .avalanchego/staking.