Releases: graphprotocol/graph-node
v0.24.2
This release only adds a fix for an issue where certain GraphQL queries
could lead to graph-node
running out of memory even on very large
systems. This release adds code that checks the size of GraphQL responses
as they are assembled, and can warn about large responses in the logs
resp. abort query execution based on the values of the two new environment
variables GRAPH_GRAPHQL_WARN_RESULT_SIZE
and
GRAPH_GRAPHQL_ERROR_RESULT_SIZE
. It also adds Prometheus metrics
query_result_size
and query_result_max
to track the memory consumption
of successful GraphQL queries. The unit for the two environment variables
is bytes, based on an estimate of the memory used by the result; it is best
to set them after observing the Prometheus metrics for a while to establish
what constitutes a reasonable limit for them.
We strongly recommend updating to this version as quickly as possible.
v0.24.1
Feature Management
This release supports the upcoming Spec Version 0.0.4 that enables subgraph features to be declared in the manifest and
validated during subgraph deployment
#2682
#2746.
Subgraphs using previous versions are still supported and won't be affected by this change.
New Indexer GraphQL query: subgraphFeatures
It is now possible to query for the features a subgraph uses given its Qm-hash ID.
For instance, the following query...
{
subgraphFeatures(subgraphId: "QmW9ajg2oTyPfdWKyUkxc7cTJejwdyCbRrSivfryTfFe5D") {
features
errors
}
}
... would produce this result:
{
"data": {
"subgraphFeatures": {
"errors": [],
"features": [
"nonFatalErrors",
"ipfsOnEthereumContracts"
]
}
}
}
Subraphs with any Spec Version can be queried that way.
Api Version 0.0.5
- Added better error message for null pointers in the runtime #2780.
Environment Variables
- When
GETH_ETH_CALL_ERRORS_ENV
is unset, it doesn't makeeth_call
errors to be considered determinsistic anymore #2784
Robustness
- Tolerate a non-primary shard being down during startup #2727.
- Check that at least one replica for each shard has a non-zero weight #2749.
- Reduce locking for the chain head listener #2763.
Logs
- Improve block ingestor error reporting for missing receipts #2743.
v0.24.0
0.24
Api Version 0.0.5
This release ships support for API version 0.0.5 in mappings. It contains a fix for call handlers
and the long awaited AssemblyScript version upgrade!
- AssemblyScript upgrade: The mapping runtime is updated to support up-to-date versions of the
AssemblyScript compiler. The graph-cli/-ts releases to support this are in alpha, soon they will
be released along with a migration guide for subgraphs. - Call handlers fix: Call handlers will never be triggered on transactions with a failed status,
resolving issue #2409. Done in #2511.
Logs
- The log
"Skipping handler because the event parameters do not match the event signature."
was downgraded from info to trace level. - Some block ingestor error logs were upgrded from debug to info level #2666.
Metrics
query_semaphore_wait_ms
is now by shard, and has thepool
andshard
labels.deployment_failed
metric added, it is1
if the subgraph has failed and0
otherwise.
Other
v0.23.1
v0.23.0
The Graph Node internals are being heavily refactored to prepare it for the multichain future.
In the meantime, here are the changes for this release:
- The
GRAPH_ETH_CALL_BY_NUMBER
environment variable has been removed. Graph Node requires an
Ethereum client that supports EIP-1898, which all major clients support. - Added support for IPFS versions larger than 0.4. Several changes to make
graph-node
more tolerant of slow/flaky IPFS nodes. - Added Ethereum ABI encoding and decoding functionality #2348.
- Experimental support for configuration files, see the documentation here.
- Better PoI performance #2329.
- Improve grafting performance and robustness by copying in batches #2293.
- Subgraph metadata storage has been simplified and reorganized. External
tools (e.g., Grafana dashboards) that access the database directly will need to be updated. - Ordering in GraphQL queries is now truly reversible
#2214 - The
GRAPH_SQL_STATEMENT_TIMEOUT
environment variable can be used to
enforce a timeout for individual SQL queries that are run in the course of
processing a GraphQL query
#2285 - Using
ethereum.call
in mappings in globals is deprecated
Graphman
Graphman is a CLI tool to manage your subgraphs. It is now included in the Docker container
#2289. And new commands have been added:
graphman copy
can copy subgraphs across DB shards #2313.graphman rewind
to rewind a deployment to a given block #2373.graphman query
to log info about a GraphQL query #2206.graphman create
to create a subgraph name #2419.
Metrics
- The
deployment_blocks_behind
metric has been removed, and a
deployment_head
metric has been added. To see how far a deployment is
behind, use the difference betweenethereum_chain_head_number
and
deployment_head
. - The
trigger_type
label was removed from the metricdeployment_trigger_processing_duration
.
v0.22.0
Feature: Block store sharding
This release makes it possible to shard the block and call cache for chain
data across multiple independent Postgres databases. This feature is considered experimental. We
encourage users to try this out in a test environment, but do not recommend it yet for production
use. In particular, the details of how sharding is configured may change in backwards-incompatible
ways in the future.
Feature: Non-fatal errors update
Non-fatal errors (see release 0.20 for details) is documented and can now be enabled on graph-cli.
Various related bug fixes have been made #2121 #2136 #2149 #2160.
Improvements
- Add bitwise operations and string constructor to BigInt #2151.
- docker: Allow custom ethereum poll interval #2139.
- Deterministic error work in preparation for gas #2112
Bug fixes
v0.21.1
v0.21.0
Source: Release notes
Feature: Database sharding
This release makes it possible to shard subgraph storage and spread subgraph deployments, and the load coming from indexing and querying them across multiple independent Postgres databases.
This feature is considered experimenatal. We encourage users to try this out in a test environment, but do not recommend it yet for production use. In particular, the details of how sharding is configured may change in backwards-incompatible ways in the future.
Breaking change: Require a block number in proofOfIndexing
queries
This changes the proofOfIndexing
GraphQL API from
type Query {
proofOfIndexing(subgraph: String!, blockHash: Bytes!, indexer: Bytes): Bytes
}
to
type Query {
proofOfIndexing(
subgraph: String!
blockNumber: Int!
blockHash: Bytes!
indexer: Bytes
): Bytes
}
This allows the indexer agent to provide a block number and hash to be able to obtain a POI even if this block is not cached in the Ethereum blocks cache. Prior to this, the POI would be null
if this wasn't the case, even if the subgraph deployment in question was up to date, leading to the indexer missing out on indexing rewards.
Misc
- Fix non-determinism caused by not (always) correctly reverting dynamic sources when handling reorgs.
- Integrate the query cache into subscriptions to improve their performance.
- Add
graphman
crate for managing Graph Node infrastructure. - Improve query cache logging.
- Expose indexing status port (
8030
) from Docker image. - Remove support for unnecessary data sources
templates
inside subgraph data sources. They are only supported at the top level. - Avoid sending empty store events through the database.
- Fix database connection deadlocks.
- Rework the codebase to use
anyhow
instead offailure
. - Log stack trace in case of database connection timeouts, to help with root-causing.
- Fix stack overflows in GraphQL parsing.
- Disable fulltext search by default (it is nondeterministic and therefore not currently supported in the network).