This is a Graphcast CLI tailored for discontinuous participation on the Graphcast P2P network. The available message type is UpgradeIntentMessage from a subgraph owner, who can use the CLI to send a single message about subgraphs they own.
When developers publish a new version (subgraph deployment) to their subgraph, data service instability may occur while their API queries the pre-existing version. Indexers may require some time to sync a subgraph to the chainhead after they have stopped syncing the previous deployment. To decrease the upgrade friction, developers can send a message before publishing the subgraph on current deployment hash's Graphcast channel, which includes the subgraph id, corresponding new deployment hash, and the represented graph account that must be validated to be the subgraph's owner.
Indexers running the subgraph radio and listening to that channel will in turn receive the message and potentially start to offchain new deployment.
It is still at the subgraph developers' discretion to await for the indexers to sync upto chainhead, in which point they can publish the staged version without disrupting API usage. This tool provides a convinence function that allows subgraph developer to ....
To send a message on Graphcast mainnet for the subgraph deployment "QmacQnSgia4iDPWHpeY6aWxesRFdb8o5DKZUx96zZqEWrB", we would need its subgraph id "CnJMdCkW3pr619gsJVtUPAWxspALPdCMw6o7obzYBNp3", private key to the graph account or an operator of the graph account, the subgraph owner's graph account, and the new deployment hash. You can supply them as an CLI argument
cargo run -- --private-key "abcdefgabcdefgabcdefgabcdefgabcdefgabcdefgabcdefgabcdefgabcdefg1" \
--graph-account "0xe9a1cabd57700b17945fd81feefba82340d9568f" \
upgrade-presync --new-hash "QmVVfLWowm1xkqc41vcygKNwFUvpsDSMbHdHghxmDVmH9x" \
--subgraph-id "CnJMdCkW3pr619gsJVtUPAWxspALPdCMw6o7obzYBNp3"
or assuming environmental variables are used for graph-account and private-key, simply use
cargo run upgrade-presync --subgraph-id CnJMdCkW3pr619gsJVtUPAWxspALPdCMw6o7obzYBNp3 --new-hash QmVVfLWowm1xkqc41vcygKNwFUvpsDSMbHdHghxmDVmH9x
The entire process from running the binary to sending the message should take ~45 seconds. One can expect the terminal to output the following:
2023-07-31T17:56:44.783866Z INFO Creating Graphcast Agent, radio_name: "subgraph-radio", registry_subgraph: "https://api.thegraph.com/subgraphs/name/hopeyen/graphcast-registry-goerli", network_subgraph: "https://api.thegraph.com/subgraphs/name/graphprotocol/graph-network-goerli", graphcast_network: "testnet", max_retry: 5
2023-07-31T17:56:59.328490Z INFO Sent message, msg_id: "0xc6b1131e0f8302abe48057f6fc69722ab46bd4285c2c4a8a8bdca6b221dcda96"
After sending UpgradeIntentMessage, a developer can periodically check the indexing status of the new subgraph deployment at the public APIs of the indexers who actively allocates on the current version of the subgraph.
Same arguments here can be used as the argument for UpgradeIntentMessage. However, gossips are not involved in this operation and the queries are made through deterministic queries.
cargo run indexing-status --subgraph-id "0x00000444e5a1a667663b0adfd853e8efa0470698-0" --new-hash QmfDJFYaDX7BdwT6rYa8Bx71vPjTueUVDN99pdwFgysDiZ
To run unit tests for the Radio. We recommend using nextest as your test runner. Once you have it installed you can run the tests using the following commands:
cargo nextest run
We welcome and appreciate your contributions! Please see the Contributor Guide, Code Of Conduct and Security Notes for this repository.