graphql-engine/server/src-lib/Hasura/Tracing/TraceId.hs

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

130 lines
3.7 KiB
Haskell
Raw Normal View History

module Hasura.Tracing.TraceId
( -- * TraceId
TraceId,
randomTraceId,
traceIdFromBytes,
traceIdToBytes,
traceIdFromHex,
traceIdToHex,
-- * SpanId
SpanId,
randomSpanId,
spanIdFromBytes,
spanIdToBytes,
spanIdFromHex,
spanIdToHex,
)
where
import Data.Bits ((.|.))
import Data.ByteString (ByteString)
import Data.ByteString qualified as ByteString
import Data.ByteString.Base16 qualified as Base16
import Data.Serialize qualified as Serialize
import Hasura.Prelude
import System.Random.Stateful qualified as Random
--------------------------------------------------------------------------------
Rewrite `Tracing` to allow for only one `TraceT` in the entire stack. This PR is on top of #7789. ### Description This PR entirely rewrites the API of the Tracing library, to make `interpTraceT` a thing of the past. Before this change, we ran traces by sticking a `TraceT` on top of whatever we were doing. This had several major drawbacks: - we were carrying a bunch of `TraceT` across the codebase, and the entire codebase had to know about it - we needed to carry a second class constraint around (`HasReporterM`) to be able to run all of those traces - we kept having to do stack rewriting with `interpTraceT`, which went from inconvenient to horrible - we had to declare several behavioral instances on `TraceT m` This PR rewrite all of `Tracing` using a more conventional model: there is ONE `TraceT` at the bottom of the stack, and there is an associated class constraint `MonadTrace`: any part of the code that happens to satisfy `MonadTrace` is able to create new traces. We NEVER have to do stack rewriting, `interpTraceT` is gone, and `TraceT` and `Reporter` become implementation details that 99% of the code is blissfully unaware of: code that needs to do tracing only needs to declare that the monad in which it operates implements `MonadTrace`. In doing so, this PR revealed **several bugs in the codebase**: places where we were expecting to trace something, but due to the default instance of `HasReporterM IO` we would actually not do anything. This PR also splits the code of `Tracing` in more byte-sized modules, with the goal of potentially moving to `server/lib` down the line. ### Remaining work This PR is a draft; what's left to do is: - [x] make Pro compile; i haven't updated `HasuraPro/Main` yet - [x] document Tracing by writing a note that explains how to use the library, and the meaning of "reporter", "trace" and "span", as well as the pitfalls - [x] discuss some of the trade-offs in the implementation, which is why i'm opening this PR already despite it not fully building yet - [x] it depends on #7789 being merged first PR-URL: https://github.com/hasura/graphql-engine-mono/pull/7791 GitOrigin-RevId: cadd32d039134c93ddbf364599a2f4dd988adea8
2023-03-13 20:37:16 +03:00
-- TraceId
-- | 128-bit trace identifiers.
--
-- 'TraceId's are guaranteed to have at least one non-zero bit.
data TraceId
= TraceId
{-# UNPACK #-} !Word64
{-# UNPACK #-} !Word64
deriving (Eq)
-- 128 bits
traceIdBytes :: Int
traceIdBytes = 16
Rewrite `Tracing` to allow for only one `TraceT` in the entire stack. This PR is on top of #7789. ### Description This PR entirely rewrites the API of the Tracing library, to make `interpTraceT` a thing of the past. Before this change, we ran traces by sticking a `TraceT` on top of whatever we were doing. This had several major drawbacks: - we were carrying a bunch of `TraceT` across the codebase, and the entire codebase had to know about it - we needed to carry a second class constraint around (`HasReporterM`) to be able to run all of those traces - we kept having to do stack rewriting with `interpTraceT`, which went from inconvenient to horrible - we had to declare several behavioral instances on `TraceT m` This PR rewrite all of `Tracing` using a more conventional model: there is ONE `TraceT` at the bottom of the stack, and there is an associated class constraint `MonadTrace`: any part of the code that happens to satisfy `MonadTrace` is able to create new traces. We NEVER have to do stack rewriting, `interpTraceT` is gone, and `TraceT` and `Reporter` become implementation details that 99% of the code is blissfully unaware of: code that needs to do tracing only needs to declare that the monad in which it operates implements `MonadTrace`. In doing so, this PR revealed **several bugs in the codebase**: places where we were expecting to trace something, but due to the default instance of `HasReporterM IO` we would actually not do anything. This PR also splits the code of `Tracing` in more byte-sized modules, with the goal of potentially moving to `server/lib` down the line. ### Remaining work This PR is a draft; what's left to do is: - [x] make Pro compile; i haven't updated `HasuraPro/Main` yet - [x] document Tracing by writing a note that explains how to use the library, and the meaning of "reporter", "trace" and "span", as well as the pitfalls - [x] discuss some of the trade-offs in the implementation, which is why i'm opening this PR already despite it not fully building yet - [x] it depends on #7789 being merged first PR-URL: https://github.com/hasura/graphql-engine-mono/pull/7791 GitOrigin-RevId: cadd32d039134c93ddbf364599a2f4dd988adea8
2023-03-13 20:37:16 +03:00
randomTraceId :: MonadIO m => m TraceId
randomTraceId = liftIO do
(w1, w2) <-
flip Random.applyAtomicGen Random.globalStdGen $ \gen0 ->
let (!w1, !gen1) = Random.random gen0
(!w2, !gen2) = Random.random gen1
in ((w1, w2), gen2)
if w1 .|. w2 == 0
then randomTraceId
else pure $ TraceId w1 w2
-- | Create a 'TraceId' from a 'ByteString'.
--
-- Fails if the 'ByteString' is not exactly 16 bytes long, or if it contains
-- only zero bytes.
traceIdFromBytes :: ByteString -> Maybe TraceId
traceIdFromBytes bs = do
guard $ ByteString.length bs == traceIdBytes
(w1, w2) <-
eitherToMaybe $
flip Serialize.runGet bs $
(,) <$> Serialize.getWord64be <*> Serialize.getWord64be
guard $ w1 .|. w2 /= 0
pure $ TraceId w1 w2
-- | Convert a 'TraceId' to a 'ByteString' of 16 bytes.
traceIdToBytes :: TraceId -> ByteString
traceIdToBytes (TraceId w1 w2) =
Serialize.runPut $ Serialize.putWord64be w1 >> Serialize.putWord64be w2
-- | Create a 'TraceId' from a 'ByteString' of hex characters.
--
-- Fails if the 'ByteString' is not exactly 32 characters long, or if it
-- contains only zero characters.
traceIdFromHex :: ByteString -> Maybe TraceId
traceIdFromHex = traceIdFromBytes <=< eitherToMaybe . Base16.decode
-- | Convert a 'TraceId' to a 'ByteString' of 32 lowercase hex characters.
traceIdToHex :: TraceId -> ByteString
traceIdToHex = Base16.encode . traceIdToBytes
--------------------------------------------------------------------------------
Rewrite `Tracing` to allow for only one `TraceT` in the entire stack. This PR is on top of #7789. ### Description This PR entirely rewrites the API of the Tracing library, to make `interpTraceT` a thing of the past. Before this change, we ran traces by sticking a `TraceT` on top of whatever we were doing. This had several major drawbacks: - we were carrying a bunch of `TraceT` across the codebase, and the entire codebase had to know about it - we needed to carry a second class constraint around (`HasReporterM`) to be able to run all of those traces - we kept having to do stack rewriting with `interpTraceT`, which went from inconvenient to horrible - we had to declare several behavioral instances on `TraceT m` This PR rewrite all of `Tracing` using a more conventional model: there is ONE `TraceT` at the bottom of the stack, and there is an associated class constraint `MonadTrace`: any part of the code that happens to satisfy `MonadTrace` is able to create new traces. We NEVER have to do stack rewriting, `interpTraceT` is gone, and `TraceT` and `Reporter` become implementation details that 99% of the code is blissfully unaware of: code that needs to do tracing only needs to declare that the monad in which it operates implements `MonadTrace`. In doing so, this PR revealed **several bugs in the codebase**: places where we were expecting to trace something, but due to the default instance of `HasReporterM IO` we would actually not do anything. This PR also splits the code of `Tracing` in more byte-sized modules, with the goal of potentially moving to `server/lib` down the line. ### Remaining work This PR is a draft; what's left to do is: - [x] make Pro compile; i haven't updated `HasuraPro/Main` yet - [x] document Tracing by writing a note that explains how to use the library, and the meaning of "reporter", "trace" and "span", as well as the pitfalls - [x] discuss some of the trade-offs in the implementation, which is why i'm opening this PR already despite it not fully building yet - [x] it depends on #7789 being merged first PR-URL: https://github.com/hasura/graphql-engine-mono/pull/7791 GitOrigin-RevId: cadd32d039134c93ddbf364599a2f4dd988adea8
2023-03-13 20:37:16 +03:00
-- SpanId
-- | 64-bit span identifiers
--
-- 'SpanId's are guaranteed to have at least one non-zero bit.
newtype SpanId = SpanId Word64
deriving (Eq)
-- 64 bits
spanIdBytes :: Int
spanIdBytes = 8
Rewrite `Tracing` to allow for only one `TraceT` in the entire stack. This PR is on top of #7789. ### Description This PR entirely rewrites the API of the Tracing library, to make `interpTraceT` a thing of the past. Before this change, we ran traces by sticking a `TraceT` on top of whatever we were doing. This had several major drawbacks: - we were carrying a bunch of `TraceT` across the codebase, and the entire codebase had to know about it - we needed to carry a second class constraint around (`HasReporterM`) to be able to run all of those traces - we kept having to do stack rewriting with `interpTraceT`, which went from inconvenient to horrible - we had to declare several behavioral instances on `TraceT m` This PR rewrite all of `Tracing` using a more conventional model: there is ONE `TraceT` at the bottom of the stack, and there is an associated class constraint `MonadTrace`: any part of the code that happens to satisfy `MonadTrace` is able to create new traces. We NEVER have to do stack rewriting, `interpTraceT` is gone, and `TraceT` and `Reporter` become implementation details that 99% of the code is blissfully unaware of: code that needs to do tracing only needs to declare that the monad in which it operates implements `MonadTrace`. In doing so, this PR revealed **several bugs in the codebase**: places where we were expecting to trace something, but due to the default instance of `HasReporterM IO` we would actually not do anything. This PR also splits the code of `Tracing` in more byte-sized modules, with the goal of potentially moving to `server/lib` down the line. ### Remaining work This PR is a draft; what's left to do is: - [x] make Pro compile; i haven't updated `HasuraPro/Main` yet - [x] document Tracing by writing a note that explains how to use the library, and the meaning of "reporter", "trace" and "span", as well as the pitfalls - [x] discuss some of the trade-offs in the implementation, which is why i'm opening this PR already despite it not fully building yet - [x] it depends on #7789 being merged first PR-URL: https://github.com/hasura/graphql-engine-mono/pull/7791 GitOrigin-RevId: cadd32d039134c93ddbf364599a2f4dd988adea8
2023-03-13 20:37:16 +03:00
randomSpanId :: MonadIO m => m SpanId
randomSpanId = liftIO do
w <- Random.uniformM Random.globalStdGen
if w == 0
then randomSpanId
else pure $ SpanId w
-- | Create a 'SpanId' from a 'ByteString'.
--
-- Fails if the 'ByteString' is not exactly 8 bytes long, or if it contains
-- only zero bytes.
spanIdFromBytes :: ByteString -> Maybe SpanId
spanIdFromBytes bs = do
guard $ ByteString.length bs == spanIdBytes
w <- eitherToMaybe $ Serialize.runGet Serialize.getWord64be bs
guard $ w /= 0
pure $ SpanId w
-- | Convert a 'SpanId' to a 'ByteString' of 8 bytes.
spanIdToBytes :: SpanId -> ByteString
spanIdToBytes (SpanId w) = Serialize.runPut $ Serialize.putWord64be w
-- | Create a 'SpanId' from a 'ByteString' of hex characters.
--
-- Fails if the 'ByteString' is not exactly 16 characters long, or if it
-- contains only zero characters.
spanIdFromHex :: ByteString -> Maybe SpanId
spanIdFromHex = spanIdFromBytes <=< eitherToMaybe . Base16.decode
-- | Convert a 'SpanId' to a 'ByteString' of 16 lowercase hex characters.
spanIdToHex :: SpanId -> ByteString
spanIdToHex = Base16.encode . spanIdToBytes