graphql-engine/server/src-lib/Hasura/GraphQL/Explain.hs
Auke Booij 7bead93827 server: remove remnants of query plan caching (fix #1795)
Query plan caching was introduced by - I believe - hasura/graphql-engine#1934 in order to reduce the query response latency. During the development of PDV in hasura/graphql-engine#4111, it was found out that the new architecture (for which query plan caching wasn't implemented) performed comparably to the pre-PDV architecture with caching. Hence, it was decided to leave query plan caching until some day in the future when it was deemed necessary.

Well, we're in the future now, and there still isn't a convincing argument for query plan caching. So the time has come to remove some references to query plan caching from the codebase. For the most part, any code being removed would probably not be very well suited to the post-PDV architecture of query execution, so arguably not much is lost.

Apart from simplifying the code, this PR will contribute towards making the GraphQL schema generation more modular, testable, and easier to profile. I'd like to eventually work towards a situation in which it's easy to generate a GraphQL schema parser *in isolation*, without being connected to a database, and then parse a GraphQL query *in isolation*, without even listening any HTTP port. It is important that both of these operations can be examined in detail, and in isolation, since they are two major performance bottlenecks, as well as phases where many important upcoming features hook into.

Implementation

The following have been removed:
- The entirety of `server/src-lib/Hasura/GraphQL/Execute/Plan.hs`
- The core phases of query parsing and execution no longer have any references to query plan caching. Note that this is not to be confused with query *response* caching, which is not affected by this PR. This includes removal of the types:
- - `Opaque`, which is replaced by a tuple. Note that the old implementation was broken and did not adequately hide the constructors.
- - `QueryReusability` (and the `markNotReusable` method). Notably, the implementation of the `ParseT` monad now consists of two, rather than three, monad transformers.
- Cache-related tests (in `server/src-test/Hasura/CacheBoundedSpec.hs`) have been removed .
- References to query plan caching in the documentation.
- The `planCacheOptions` in the `TenantConfig` type class was removed. However, during parsing, unrecognized fields in the YAML config get ignored, so this does not cause a breaking change. (Confirmed manually, as well as in consultation with @sordina.)
- The metrics no longer send cache hit/miss messages.

There are a few places in which one can still find references to query plan caching:

- We still accept the `--query-plan-cache-size` command-line option for backwards compatibility. The `HASURA_QUERY_PLAN_CACHE_SIZE` environment variable is not read.

https://github.com/hasura/graphql-engine-mono/pull/1815

GitOrigin-RevId: 17d92b254ec093c62a7dfeec478658ede0813eb7
2021-07-27 11:52:43 +00:00

122 lines
5.3 KiB
Haskell

module Hasura.GraphQL.Explain
( explainGQLQuery
, GQLExplain
) where
import Hasura.Prelude
import qualified Data.Aeson as J
import qualified Data.Aeson.TH as J
import qualified Data.HashMap.Strict as Map
import qualified Data.HashMap.Strict.InsOrd as OMap
import qualified Language.GraphQL.Draft.Syntax as G
import Control.Monad.Trans.Control (MonadBaseControl)
import qualified Hasura.GraphQL.Execute as E
import qualified Hasura.GraphQL.Execute.Action as E
import qualified Hasura.GraphQL.Execute.Inline as E
import qualified Hasura.GraphQL.Execute.Query as E
import qualified Hasura.GraphQL.Execute.RemoteJoin.Collect as RJ
import qualified Hasura.GraphQL.Transport.HTTP.Protocol as GH
import qualified Hasura.SQL.AnyBackend as AB
import Hasura.Base.Error
import Hasura.EncJSON
import Hasura.GraphQL.Execute.Backend
import Hasura.GraphQL.Execute.Instances ()
import Hasura.GraphQL.Parser
import Hasura.GraphQL.Transport.Backend
import Hasura.GraphQL.Transport.Instances ()
import Hasura.Metadata.Class
import Hasura.RQL.IR
import Hasura.RQL.Types
import Hasura.Session
data GQLExplain
= GQLExplain
{ _gqeQuery :: !GH.GQLReqParsed
, _gqeUser :: !(Maybe (Map.HashMap Text Text))
, _gqeIsRelay :: !(Maybe Bool)
} deriving (Show, Eq)
$(J.deriveJSON hasuraJSON{J.omitNothingFields=True}
''GQLExplain
)
-- NOTE: This function has a 'MonadTrace' constraint in master, but we don't need it
-- here. We should evaluate if we need it here.
explainQueryField
:: ( MonadError QErr m
, MonadIO m
)
=> UserInfo
-> G.Name
-> QueryRootField UnpreparedValue
-> m EncJSON
explainQueryField userInfo fieldName rootField = do
case rootField of
RFRemote _ -> throw400 InvalidParams "only hasura queries can be explained"
RFAction _ -> throw400 InvalidParams "query actions cannot be explained"
RFRaw _ -> pure $ encJFromJValue $ ExplainPlan fieldName Nothing Nothing
RFDB sourceName exists -> do
step <- AB.dispatchAnyBackend @BackendExecute exists
\(SourceConfigWith sourceConfig (QDBR db)) -> do
let (newDB, remoteJoins) = RJ.getRemoteJoins db
unless (isNothing remoteJoins) $
throw400 InvalidParams "queries with remote relationships cannot be explained"
mkDBQueryExplain fieldName userInfo sourceName sourceConfig newDB
AB.dispatchAnyBackend @BackendTransport step runDBQueryExplain
explainGQLQuery
:: forall m
. ( MonadError QErr m
, MonadIO m
, MonadBaseControl IO m
, MonadMetadataStorage (MetadataStorageT m)
)
=> SchemaCache
-> GQLExplain
-> m EncJSON
explainGQLQuery sc (GQLExplain query userVarsRaw maybeIsRelay) = do
-- NOTE!: we will be executing what follows as though admin role. See e.g. notes in explainField:
userInfo <-
mkUserInfo (URBFromSessionVariablesFallback adminRoleName) UAdminSecretSent
sessionVariables
-- we don't need to check in allow list as we consider it an admin endpoint
let takeFragment =
\case G.ExecutableDefinitionFragment f -> Just f; _ -> Nothing
fragments = mapMaybe takeFragment $ GH.unGQLExecDoc $ GH._grQuery query
(graphQLContext, queryParts) <- E.getExecPlanPartial userInfo sc queryType query
case queryParts of
G.TypedOperationDefinition G.OperationTypeQuery _ varDefs directives selSet -> do
-- (Here the above fragment inlining is actually executed.)
inlinedSelSet <- E.inlineSelectionSet fragments selSet
(unpreparedQueries, _, _) <-
E.parseGraphQLQuery graphQLContext varDefs (GH._grVariables query) directives inlinedSelSet
-- TODO: validate directives here
encJFromList <$>
for (OMap.toList unpreparedQueries) (uncurry (explainQueryField userInfo))
G.TypedOperationDefinition G.OperationTypeMutation _ _ _ _ ->
throw400 InvalidParams "only queries can be explained"
G.TypedOperationDefinition G.OperationTypeSubscription _ varDefs directives selSet -> do
-- (Here the above fragment inlining is actually executed.)
inlinedSelSet <- E.inlineSelectionSet fragments selSet
(unpreparedQueries, _, _) <- E.parseGraphQLQuery graphQLContext varDefs (GH._grVariables query) directives inlinedSelSet
-- TODO: validate directives here
validSubscription <- E.buildSubscriptionPlan userInfo unpreparedQueries
case validSubscription of
E.SEAsyncActionsWithNoRelationships _ -> throw400 NotSupported "async action query fields without relationships to table cannot be explained"
E.SEOnSourceDB actionIds liveQueryBuilder -> do
actionLogResponseMap <- fst <$> E.fetchActionLogResponses actionIds
(_, E.LQP exists) <- liftEitherM $ liftIO $ runExceptT $ liveQueryBuilder actionLogResponseMap
AB.dispatchAnyBackend @BackendExecute exists \(E.MultiplexedLiveQueryPlan execPlan) ->
encJFromJValue <$> mkLiveQueryExplain execPlan
where
queryType = bool E.QueryHasura E.QueryRelay $ Just True == maybeIsRelay
sessionVariables = mkSessionVariablesText $ fromMaybe mempty userVarsRaw