This is the first PR towards filesystem support for file:/// URLs
compatible with #11264. Currently it plumbs in only LDS filesystem
support for list collections with only inline entries.
Risk level: Low (opt in)
Testing: Unit and integration tests added.
Signed-off-by: Harvey Tuch <htuch@google.com>
Mirrored from https://github.com/envoyproxy/envoy @ 108d2bcc884a0249eda26213e5dbdd0124134558
Commit Message: CacheFilter: parses the allowed_vary_headers from the cache config.
Additional Description:
Parses the allowlist from the cache config proto; this allows users to define a set of rules to control which headers can be varied in the cache.
Risk Level: Low
Testing: Unit testing
Docs Changes: Updated cache proto's comments regarding allowed_vary_headers
Release Notes: N/A
Fixes#10131
Signed-off-by: Caio <caiomelo@google.com>
Mirrored from https://github.com/envoyproxy/envoy @ 6a994d5afcd7fa038beac5113951afe6daaaf677
Tweak the HTTP upgrades documentation to mention RFC8841 in the
documentation body (complementing the existing link to the RFC).
Minor fix to the warning text for CONNECT support.
Make explicit mention of "Extended CONNECT" in the API docs for
`RouteMatch`.
Closes#13044.
Signed-off-by: Nick Travers <n.e.travers@gmail.com>
Mirrored from https://github.com/envoyproxy/envoy @ 12d7b17906df69749990a9f481bc4fce087c22e6
This patch introduces the xDS transport++ context parameter encoding
algorithm.
Risk level: Low
Testing: Unit tests added.
Part of #11264
Signed-off-by: Harvey Tuch <htuch@google.com>
Mirrored from https://github.com/envoyproxy/envoy @ 8613b8ebbb2e9465177eb1d437f28363b92225c3
Add a field to the current protobuf of ScopedRouteConfiguration to enable on demand scoped route table loading. The on demand scope route tables will be loaded lazily. The lazy loading feature of route table associated with scope is achieved by extending the current vhds on_demand filter to support lazy loading of RouteConfigurationscoped route discovery service.If a scoped route configuration is set to be loaded lazily, upon a http request using SRDS, when the corresponding route table of a scope is not found, post a callback to control plane, request the route table from the management server, after the route table has been initialized, continue the filter chain.
https://docs.google.com/document/d/15GX30U5CH2bsWUyQRkiiQ_nbMCoklvgP_ObrDaSlkuc/edit?usp=sharing
Risk Level: Low
Testing: add unit tests and integration test to verifiy behavior changes
Fixes#10641
Signed-off-by: chaoqinli <chaoqinli@google.com>
Mirrored from https://github.com/envoyproxy/envoy @ 709d1c31749a6ba2eab2865927f10300570ac533
This introduces an option to entirely omit null values from the access log.
Risk Level: Low
Testing: Unit and integration tests
Docs Changes: New option documented in proto file
Release Notes: Updated
Fixes#12735
Signed-off-by: Petr Pchelko <ppchelko@wikimedia.org>
Mirrored from https://github.com/envoyproxy/envoy @ 9d466c71ab217317d3e989b261eb496877348a47
Currently, the maglev hash algorithm default to table size to 65537.
It is the recommended size by a paper but it is better if the user
can set this value.
This patch introduces a new MaglevLbConfig that contains table
size of maglev.
So, now, the user can set the table size of maglev by their situation.
Signed-off-by: DongRyeol Cha <dr83.cha@samsung.com>
Mirrored from https://github.com/envoyproxy/envoy @ 5fd73ca889aa12618c626c96bb33fde4707f8bf0
This can be used in similar situations to maglev/ring, but can be used when the host selection must remain the same for existing sessions even if more hosts are added to the cluster.
In a test with 1500 endpoints, updating 50 (remove 50, add 50 different ones) took:
without new setting: ~500ms
with new setting (this PR): ~4ms
Signed-off-by: Greg Greenway <ggreenway@apple.com>
Mirrored from https://github.com/envoyproxy/envoy @ 724df08c329c8270509e2a8568be90fd43d597c1
opt-in for serving requests/responses with Content-Length and Transfer-Encoding: chunked. Per RFC remove Content-Length header before forwarding it to upstream.
Signed-off-by: Oleg Guba <oleg@dropbox.com>
Mirrored from https://github.com/envoyproxy/envoy @ 954c93c28dfd6f152bed1ab81534c79faf154490
This is as described in the paper https://arxiv.org/abs/1608.01350; Logic followed is similar to what's in haproxy.
Signed-off-by: Santosh Cheler <santosh.cheler@appdynamics.com>
Mirrored from https://github.com/envoyproxy/envoy @ 760a164640acf75a8009ac89b5ff06f78a733221
Add connect_pool_per_downstream_connection flag to the cluster config (disabled by default)
Add a hashKey method to Connection in order to be able to hash on the connection ID
Signed-off-by: Craig Radcliffe <craig.radcliffe@broadcom.com>
Mirrored from https://github.com/envoyproxy/envoy @ 76a1b24dd511bb5b3a38da3e84c9003ccadc64ce
In the closed-source implementation of the AWS X-Ray tracer extension
there was additional span metadata that was hard-coded for the AWS App Mesh
service. This was expectedly removed from the implementation that was
upstreamed but there was a miss in adding an API to re-add this metadata
to spans.
Signed-off-by: Scott LaVigne <lavignes@amazon.com>
Mirrored from https://github.com/envoyproxy/envoy @ 64731cece4fb1c7a63251007e3f8bf319d1cb895
For matching values in the header that might be somewhere in the middle of the header, the present option is to use Regex in the form .Search-Pattern.. This can cause catastrophic backtracking as described in #7728
As a solution, I have introduced another header match type called contains which is based on absl::StrContains().
Risk Level: Low
Testing: Unit tests are included and manual testing was performed.
Fixes#12590
Signed-off-by: Shivanshu Goswami <shigoswami@ebay.com>
Mirrored from https://github.com/envoyproxy/envoy @ e322daaf0ca70fd3fcb1c1405830c73395510d93
This was added for gRPC server support, but we've decided to use resource names instead to explicitly request the listeners we want by name. This is more in-line with the new naming scheme described in the "xDS Transport Next Steps" design.
Signed-off-by: Mark D. Roth <roth@google.com>
Mirrored from https://github.com/envoyproxy/envoy @ 05cbb309b828dc86737c51fd2c79d30e48e397a4
This PR implements a new retry back off strategy that uses values from response headers like Retry-After or X-RateLimit-Reset (the headers are configurable) to decide the back off interval before retrying a request.
Signed-off-by: Martin Matusiak <numerodix@gmail.com>
Mirrored from https://github.com/envoyproxy/envoy @ bd2b989c578b2472faaff44902573e5b187f671f
Establish an extension point for actions to run based on Watch Dog Events.
Signed-off-by: Kevin Baichoo <kbaichoo@google.com>
Mirrored from https://github.com/envoyproxy/envoy @ 8c312f2c5b40b5ce7f6f68a1f9d4b0e98ef3829e
Change the Overload Manager API to extend the overload action state with non-binary values. This will allow future overload actions to take effect in response to increasing load, instead of to the existing inactive/active binary values. This PR also adds a range field to the overload trigger config, though no actions currently trigger on the 'scaling' state.
The existing behavior is preserved by replacing all places where OverloadActionState::Active was being used with OverloadActionState::saturated(). This is a minimal re-hashing of #11697 for ##11427.
Risk Level: medium
Testing: ran unit and integration tests
Docs Changes: none
Release Notes: add "scaling" trigger for OverloadManager actions
Signed-off-by: Alex Konradi <akonradi@google.com>
Mirrored from https://github.com/envoyproxy/envoy @ 28df333572e71fc6850e82437a59e9e378d7b00b
This implements a host_rewrite_path option for rewriting the Host header based on path. See rational in the linked issue.
Note: the regex is executed on the path with query/fragment stripped. This is analogues to what regex_rewrite option is doing.
Risk Level: Low
Testing: added unit tests
Docs Changes: document the new option in proto file
Release Notes: added to current.rst
Fixes#12430
Signed-off-by: Petr Pchelko <ppchelko@wikimedia.org>
Mirrored from https://github.com/envoyproxy/envoy @ 374dca7905fc048be74169a7655d0462606555ad
The log action will be used to set the dynamic metadata key "envoy.log", which can be used to decide whether to log a request.
Signed-off-by: davidraskin <draskin@google.com>
Mirrored from https://github.com/envoyproxy/envoy @ 89b594e09e2f4ebaac43c6b38d778a3e0dc13a3a
This is the 1st PR for #11832 that factors out the TAP filter matcher to prepare for reuse in other filters.
Signed-off-by: Yangmin Zhu <ymzhu@google.com>
Mirrored from https://github.com/envoyproxy/envoy @ 520389e677cdcd4a85df769deb40f6cdd2f4f6f8
Introduces UdpPacketWriter Interface that can be used to perform writes in Batched/PassThrough modes by using QuicGsoBatchWriter implementation from QUICHE extension.
**Additional Description:**
UDP GSO (Generic Segmentation Offload) was introduced in Linux at version 4.18. It allows batch-writing of multiple messages into a single payload and sending these messages along as a batch in a single sendmsg syscall.
Currently, Envoy performs the sending of messages using simple sendmsg implementation in pass-through mode, i.e. no support for batch writing. With this change, UdpListener can use UdpPacketWriter interface as a DefaultWriter or a GsoBatchWriter to perform pass-through or batched writes respectively.
Detailed description of the changes can be found in the design document, [here](https://docs.google.com/document/d/16ePbgkfrzQ6v-cOVMSnKDja3dUdZvX-mxT9jw29rx4g/edit?usp=sharing).
**Risk Level:** Low, not in use
**Testing:**
- Added udp_listener_impl_batched_writes_test, to verify that multiple packets of varying sizes are batched/flushed as per gso specifications while using UdpGsoBatchWriter.
- Modified existing tests, to verify that UdpDefaultWriter performs writes in pass-through mode.
- Ran all tests. All 677 tests passed successfully.
```
**Docs Changes:** None
**Release Notes:** None
**Fixes:** #11925
Signed-off-by: Yugant <yugant@google.com>
Mirrored from https://github.com/envoyproxy/envoy @ 933e267db60961e76549ae6e65ac4872cb78db4d
Add authority field in envoy grpc message to override the default host name as cluster name.
Risk Level: Low
Testing:
Docs Changes:
Release Notes:
Fix#12116
Signed-off-by: Yuchen Dai <silentdai@gmail.com>
Mirrored from https://github.com/envoyproxy/envoy @ 29b49276db0812b566d143a79996848819248428
Adding a filter for access logs that will decide whether to log based on dynamic metadata.
Signed-off-by: davidraskin <draskin@google.com>
Mirrored from https://github.com/envoyproxy/envoy @ d7c7e9a79eed80afa56a0b05cf6adf0516750d14
Implementing per-upstream prefetching, useful for high QPS or latency-sensitive services.
Risk Level: low (off by default)
Testing: new unit tests, integration test
Docs Changes: n/a
Release Notes: not yet
Signed-off-by: Alyssa Wilk <alyssar@chromium.org>
Mirrored from https://github.com/envoyproxy/envoy @ f6e90f2966887b25a8b6d5dd7f13ffb32865d66b
This ensures that the configuration is valid.
Note that this adds stricter validation to an existing proto field, but the field was only added 5 days ago, so only someone running HEAD, using this new config, with invalid/non-sensical config should be affected.
Signed-off-by: Greg Greenway <ggreenway@apple.com>
Mirrored from https://github.com/envoyproxy/envoy @ 61c2816f704d11f66a507b76abc8f01c1816fe57
Previously, a hard-coded set of buckets were used, which were not appropriate for the range or required precision of some histograms.
Fixes#7599
Signed-off-by: Greg Greenway <ggreenway@apple.com>
Mirrored from https://github.com/envoyproxy/envoy @ 0a0fc51e25599665d772f274c54addf3891ef5c7
Created a new struct for optional cluster stats. Moved timeout budget stats and added request response headers and body stats in the new struct.
Risk Level: Low
Testing: Added test cases
Docs Changes: added
Release Notes: added
Fixes#10308 , Fixes#3621
Signed-off-by: Ranjith Kumar <ranjith.dakshana2015@gmail.com>
Mirrored from https://github.com/envoyproxy/envoy @ a24c95e4fe035ae5d05691ee2b2bd7011093a9e9
Optional buffer on statsd udp
Signed-off-by: John Murray <murray@stripe.com>
Mirrored from https://github.com/envoyproxy/envoy @ 90c6a27bb28b4f8288392a53af8d2e88ef059a8b
WatchDog will now kill if max(2, registered_threads * multi_kill_threshold) threads have gone above the multikill_timeout.
Signed-off-by: Kevin Baichoo <kbaichoo@google.com>
Mirrored from https://github.com/envoyproxy/envoy @ 7f78581116ecdc9dcca319ebe68d4c8ac1d817ba
* fix filter fuzzer crash from OSS fuzz by checking for non-implemented proto fields
This will raise an exception that will be caught and abort the test run when an unimplemented oneof field is hit by the fuzzer
Signed-off-by: Sam Flattery <samflattery@google.com>
Mirrored from https://github.com/envoyproxy/envoy @ e6c57fab8022d9df228052f1f517a57584030654
Pulling out version to a separate package
Risk Level: Low
Signed-off-by: Lizan Zhou <lizan@tetrate.io>
Mirrored from https://github.com/envoyproxy/envoy @ 9a5de98bf26280a3b57882413dfb9084c5dbb2a6
This unifies HTTP/1.1 and HTTP/2 stream error on invalid messaging. Previously HTTP/1.1 defaulted permissive and HTTP/2 defaulted to strict. This defaults both to strict, resetting connections on invalid requests. This will have a major latency impact if downstream is sending a mix of valid and invalid requests over HTTP/1.1
Additional Description:
This change is runtime guarded per default behavioral change rules. It can also be reverted by setting the default to permissive (for prior HTTP/1 behvior) then overriding HTTP/2 to struct (for prior HTTP/2 behavior).
This works in conjunction with #11714, as the HTTP connection manager enforces the strictness, so the responses need to be sent via the HTTP connection manager to have strictness applied correctly.
Risk Level: High (HCM changes)
Testing: new unit tests, updated integration tests
Docs Changes: n/a
Release Notes: inline
Runtime guard: envoy.reloadable_features.hcm_stream_error_on_invalid_message
Fixes#9846
Signed-off-by: Alyssa Wilk <alyssar@chromium.org>
Mirrored from https://github.com/envoyproxy/envoy @ 88dcb292817946510bb87f8f379a954962cece79
data-plane-api(CircleCI)