forgejo/services/webhook/deliver.go

259 lines
7.4 KiB
Go
Raw Normal View History

// Copyright 2019 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
package webhook
import (
"context"
"crypto/tls"
"fmt"
"io"
"net/http"
"net/url"
"strings"
"sync"
"time"
webhook_model "code.gitea.io/gitea/models/webhook"
"code.gitea.io/gitea/modules/graceful"
"code.gitea.io/gitea/modules/hostmatcher"
"code.gitea.io/gitea/modules/log"
"code.gitea.io/gitea/modules/process"
"code.gitea.io/gitea/modules/proxy"
"code.gitea.io/gitea/modules/queue"
"code.gitea.io/gitea/modules/setting"
"code.gitea.io/gitea/modules/timeutil"
webhook_module "code.gitea.io/gitea/modules/webhook"
"github.com/gobwas/glob"
)
Store webhook event in database (#29145) Refactor the webhook logic, to have the type-dependent processing happen only in one place. --- 1. An event happens 2. It is pre-processed (depending on the webhook type) and its body is added to a task queue 3. When the task is processed, some more logic (depending on the webhook type as well) is applied to make an HTTP request This means that webhook-type dependant logic is needed in step 2 and 3. This is cumbersome and brittle to maintain. Updated webhook flow with this PR: 1. An event happens 2. It is stored as-is and added to a task queue 3. When the task is processed, the event is processed (depending on the webhook type) to make an HTTP request So the only webhook-type dependent logic happens in one place (step 3) which should be much more robust. - the raw event must be stored in the hooktask (until now, the pre-processed body was stored) - to ensure that previous hooktasks are correctly sent, a `payload_version` is added (version 1: the body has already been pre-process / version 2: the body is the raw event) So future webhook additions will only have to deal with creating an http.Request based on the raw event (no need to adjust the code in multiple places, like currently). Moreover since this processing happens when fetching from the task queue, it ensures that the queuing of new events (upon a `git push` for instance) does not get slowed down by a slow webhook. As a concrete example, the PR #19307 for custom webhooks, should be substantially smaller: - no need to change `services/webhook/deliver.go` - minimal change in `services/webhook/webhook.go` (add the new webhook to the map) - no need to change all the individual webhook files (since with this refactor the `*webhook_model.Webhook` is provided as argument) (cherry picked from commit 26653b196bd1d15c532af41f60351596dd4330bd) Conflicts: services/webhook/deliver_test.go trivial context conflict
2024-03-07 23:18:38 +01:00
// Deliver creates the [http.Request] (depending on the webhook type), sends it
// and records the status and response.
func Deliver(ctx context.Context, t *webhook_model.HookTask) error {
w, err := webhook_model.GetWebhookByID(ctx, t.HookID)
Add Webhook authorization header (#20926) _This is a different approach to #20267, I took the liberty of adapting some parts, see below_ ## Context In some cases, a weebhook endpoint requires some kind of authentication. The usual way is by sending a static `Authorization` header, with a given token. For instance: - Matrix expects a `Bearer <token>` (already implemented, by storing the header cleartext in the metadata - which is buggy on retry #19872) - TeamCity #18667 - Gitea instances #20267 - SourceHut https://man.sr.ht/graphql.md#authentication-strategies (this is my actual personal need :) ## Proposed solution Add a dedicated encrypt column to the webhook table (instead of storing it as meta as proposed in #20267), so that it gets available for all present and future hook types (especially the custom ones #19307). This would also solve the buggy matrix retry #19872. As a first step, I would recommend focusing on the backend logic and improve the frontend at a later stage. For now the UI is a simple `Authorization` field (which could be later customized with `Bearer` and `Basic` switches): ![2022-08-23-142911](https://user-images.githubusercontent.com/3864879/186162483-5b721504-eef5-4932-812e-eb96a68494cc.png) The header name is hard-coded, since I couldn't fine any usecase justifying otherwise. ## Questions - What do you think of this approach? @justusbunsi @Gusted @silverwind - ~~How are the migrations generated? Do I have to manually create a new file, or is there a command for that?~~ - ~~I started adding it to the API: should I complete it or should I drop it? (I don't know how much the API is actually used)~~ ## Done as well: - add a migration for the existing matrix webhooks and remove the `Authorization` logic there _Closes #19872_ Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com> Co-authored-by: Gusted <williamzijl7@hotmail.com> Co-authored-by: delvh <dev.lh@web.de>
2022-11-03 19:23:20 +01:00
if err != nil {
return err
}
Store webhook event in database (#29145) Refactor the webhook logic, to have the type-dependent processing happen only in one place. --- 1. An event happens 2. It is pre-processed (depending on the webhook type) and its body is added to a task queue 3. When the task is processed, some more logic (depending on the webhook type as well) is applied to make an HTTP request This means that webhook-type dependant logic is needed in step 2 and 3. This is cumbersome and brittle to maintain. Updated webhook flow with this PR: 1. An event happens 2. It is stored as-is and added to a task queue 3. When the task is processed, the event is processed (depending on the webhook type) to make an HTTP request So the only webhook-type dependent logic happens in one place (step 3) which should be much more robust. - the raw event must be stored in the hooktask (until now, the pre-processed body was stored) - to ensure that previous hooktasks are correctly sent, a `payload_version` is added (version 1: the body has already been pre-process / version 2: the body is the raw event) So future webhook additions will only have to deal with creating an http.Request based on the raw event (no need to adjust the code in multiple places, like currently). Moreover since this processing happens when fetching from the task queue, it ensures that the queuing of new events (upon a `git push` for instance) does not get slowed down by a slow webhook. As a concrete example, the PR #19307 for custom webhooks, should be substantially smaller: - no need to change `services/webhook/deliver.go` - minimal change in `services/webhook/webhook.go` (add the new webhook to the map) - no need to change all the individual webhook files (since with this refactor the `*webhook_model.Webhook` is provided as argument) (cherry picked from commit 26653b196bd1d15c532af41f60351596dd4330bd) Conflicts: services/webhook/deliver_test.go trivial context conflict
2024-03-07 23:18:38 +01:00
defer func() {
err := recover()
if err == nil {
return
}
// There was a panic whilst delivering a hook...
log.Error("PANIC whilst trying to deliver webhook task[%d] to webhook %s Panic: %v\nStacktrace: %s", t.ID, w.URL, err, log.Stack(2))
}()
t.IsDelivered = true
2024-03-20 15:44:01 +01:00
handler := GetWebhookHandler(w.Type)
if handler == nil {
return fmt.Errorf("GetWebhookHandler %q", w.Type)
}
if t.PayloadVersion == 1 {
handler = defaultHandler{true}
Store webhook event in database (#29145) Refactor the webhook logic, to have the type-dependent processing happen only in one place. --- 1. An event happens 2. It is pre-processed (depending on the webhook type) and its body is added to a task queue 3. When the task is processed, some more logic (depending on the webhook type as well) is applied to make an HTTP request This means that webhook-type dependant logic is needed in step 2 and 3. This is cumbersome and brittle to maintain. Updated webhook flow with this PR: 1. An event happens 2. It is stored as-is and added to a task queue 3. When the task is processed, the event is processed (depending on the webhook type) to make an HTTP request So the only webhook-type dependent logic happens in one place (step 3) which should be much more robust. - the raw event must be stored in the hooktask (until now, the pre-processed body was stored) - to ensure that previous hooktasks are correctly sent, a `payload_version` is added (version 1: the body has already been pre-process / version 2: the body is the raw event) So future webhook additions will only have to deal with creating an http.Request based on the raw event (no need to adjust the code in multiple places, like currently). Moreover since this processing happens when fetching from the task queue, it ensures that the queuing of new events (upon a `git push` for instance) does not get slowed down by a slow webhook. As a concrete example, the PR #19307 for custom webhooks, should be substantially smaller: - no need to change `services/webhook/deliver.go` - minimal change in `services/webhook/webhook.go` (add the new webhook to the map) - no need to change all the individual webhook files (since with this refactor the `*webhook_model.Webhook` is provided as argument) (cherry picked from commit 26653b196bd1d15c532af41f60351596dd4330bd) Conflicts: services/webhook/deliver_test.go trivial context conflict
2024-03-07 23:18:38 +01:00
}
2024-03-20 15:44:01 +01:00
req, body, err := handler.NewRequest(ctx, w, t)
Store webhook event in database (#29145) Refactor the webhook logic, to have the type-dependent processing happen only in one place. --- 1. An event happens 2. It is pre-processed (depending on the webhook type) and its body is added to a task queue 3. When the task is processed, some more logic (depending on the webhook type as well) is applied to make an HTTP request This means that webhook-type dependant logic is needed in step 2 and 3. This is cumbersome and brittle to maintain. Updated webhook flow with this PR: 1. An event happens 2. It is stored as-is and added to a task queue 3. When the task is processed, the event is processed (depending on the webhook type) to make an HTTP request So the only webhook-type dependent logic happens in one place (step 3) which should be much more robust. - the raw event must be stored in the hooktask (until now, the pre-processed body was stored) - to ensure that previous hooktasks are correctly sent, a `payload_version` is added (version 1: the body has already been pre-process / version 2: the body is the raw event) So future webhook additions will only have to deal with creating an http.Request based on the raw event (no need to adjust the code in multiple places, like currently). Moreover since this processing happens when fetching from the task queue, it ensures that the queuing of new events (upon a `git push` for instance) does not get slowed down by a slow webhook. As a concrete example, the PR #19307 for custom webhooks, should be substantially smaller: - no need to change `services/webhook/deliver.go` - minimal change in `services/webhook/webhook.go` (add the new webhook to the map) - no need to change all the individual webhook files (since with this refactor the `*webhook_model.Webhook` is provided as argument) (cherry picked from commit 26653b196bd1d15c532af41f60351596dd4330bd) Conflicts: services/webhook/deliver_test.go trivial context conflict
2024-03-07 23:18:38 +01:00
if err != nil {
return fmt.Errorf("cannot create http request for webhook %s[%d %s]: %w", w.Type, w.ID, w.URL, err)
Add Webhook authorization header (#20926) _This is a different approach to #20267, I took the liberty of adapting some parts, see below_ ## Context In some cases, a weebhook endpoint requires some kind of authentication. The usual way is by sending a static `Authorization` header, with a given token. For instance: - Matrix expects a `Bearer <token>` (already implemented, by storing the header cleartext in the metadata - which is buggy on retry #19872) - TeamCity #18667 - Gitea instances #20267 - SourceHut https://man.sr.ht/graphql.md#authentication-strategies (this is my actual personal need :) ## Proposed solution Add a dedicated encrypt column to the webhook table (instead of storing it as meta as proposed in #20267), so that it gets available for all present and future hook types (especially the custom ones #19307). This would also solve the buggy matrix retry #19872. As a first step, I would recommend focusing on the backend logic and improve the frontend at a later stage. For now the UI is a simple `Authorization` field (which could be later customized with `Bearer` and `Basic` switches): ![2022-08-23-142911](https://user-images.githubusercontent.com/3864879/186162483-5b721504-eef5-4932-812e-eb96a68494cc.png) The header name is hard-coded, since I couldn't fine any usecase justifying otherwise. ## Questions - What do you think of this approach? @justusbunsi @Gusted @silverwind - ~~How are the migrations generated? Do I have to manually create a new file, or is there a command for that?~~ - ~~I started adding it to the API: should I complete it or should I drop it? (I don't know how much the API is actually used)~~ ## Done as well: - add a migration for the existing matrix webhooks and remove the `Authorization` logic there _Closes #19872_ Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com> Co-authored-by: Gusted <williamzijl7@hotmail.com> Co-authored-by: delvh <dev.lh@web.de>
2022-11-03 19:23:20 +01:00
}
// Record delivery information.
t.RequestInfo = &webhook_model.HookRequest{
URL: req.URL.String(),
HTTPMethod: req.Method,
Headers: map[string]string{},
Store webhook event in database (#29145) Refactor the webhook logic, to have the type-dependent processing happen only in one place. --- 1. An event happens 2. It is pre-processed (depending on the webhook type) and its body is added to a task queue 3. When the task is processed, some more logic (depending on the webhook type as well) is applied to make an HTTP request This means that webhook-type dependant logic is needed in step 2 and 3. This is cumbersome and brittle to maintain. Updated webhook flow with this PR: 1. An event happens 2. It is stored as-is and added to a task queue 3. When the task is processed, the event is processed (depending on the webhook type) to make an HTTP request So the only webhook-type dependent logic happens in one place (step 3) which should be much more robust. - the raw event must be stored in the hooktask (until now, the pre-processed body was stored) - to ensure that previous hooktasks are correctly sent, a `payload_version` is added (version 1: the body has already been pre-process / version 2: the body is the raw event) So future webhook additions will only have to deal with creating an http.Request based on the raw event (no need to adjust the code in multiple places, like currently). Moreover since this processing happens when fetching from the task queue, it ensures that the queuing of new events (upon a `git push` for instance) does not get slowed down by a slow webhook. As a concrete example, the PR #19307 for custom webhooks, should be substantially smaller: - no need to change `services/webhook/deliver.go` - minimal change in `services/webhook/webhook.go` (add the new webhook to the map) - no need to change all the individual webhook files (since with this refactor the `*webhook_model.Webhook` is provided as argument) (cherry picked from commit 26653b196bd1d15c532af41f60351596dd4330bd) Conflicts: services/webhook/deliver_test.go trivial context conflict
2024-03-07 23:18:38 +01:00
Body: string(body),
}
for k, vals := range req.Header {
t.RequestInfo.Headers[k] = strings.Join(vals, ",")
}
Store webhook event in database (#29145) Refactor the webhook logic, to have the type-dependent processing happen only in one place. --- 1. An event happens 2. It is pre-processed (depending on the webhook type) and its body is added to a task queue 3. When the task is processed, some more logic (depending on the webhook type as well) is applied to make an HTTP request This means that webhook-type dependant logic is needed in step 2 and 3. This is cumbersome and brittle to maintain. Updated webhook flow with this PR: 1. An event happens 2. It is stored as-is and added to a task queue 3. When the task is processed, the event is processed (depending on the webhook type) to make an HTTP request So the only webhook-type dependent logic happens in one place (step 3) which should be much more robust. - the raw event must be stored in the hooktask (until now, the pre-processed body was stored) - to ensure that previous hooktasks are correctly sent, a `payload_version` is added (version 1: the body has already been pre-process / version 2: the body is the raw event) So future webhook additions will only have to deal with creating an http.Request based on the raw event (no need to adjust the code in multiple places, like currently). Moreover since this processing happens when fetching from the task queue, it ensures that the queuing of new events (upon a `git push` for instance) does not get slowed down by a slow webhook. As a concrete example, the PR #19307 for custom webhooks, should be substantially smaller: - no need to change `services/webhook/deliver.go` - minimal change in `services/webhook/webhook.go` (add the new webhook to the map) - no need to change all the individual webhook files (since with this refactor the `*webhook_model.Webhook` is provided as argument) (cherry picked from commit 26653b196bd1d15c532af41f60351596dd4330bd) Conflicts: services/webhook/deliver_test.go trivial context conflict
2024-03-07 23:18:38 +01:00
// Add Authorization Header
authorization, err := w.HeaderAuthorization()
if err != nil {
return fmt.Errorf("cannot get Authorization header for webhook %s[%d %s]: %w", w.Type, w.ID, w.URL, err)
}
if authorization != "" {
req.Header.Set("Authorization", authorization)
redacted := "******"
if strings.HasPrefix(authorization, "Bearer ") {
redacted = "Bearer " + redacted
} else if strings.HasPrefix(authorization, "Basic ") {
redacted = "Basic " + redacted
}
t.RequestInfo.Headers["Authorization"] = redacted
Store webhook event in database (#29145) Refactor the webhook logic, to have the type-dependent processing happen only in one place. --- 1. An event happens 2. It is pre-processed (depending on the webhook type) and its body is added to a task queue 3. When the task is processed, some more logic (depending on the webhook type as well) is applied to make an HTTP request This means that webhook-type dependant logic is needed in step 2 and 3. This is cumbersome and brittle to maintain. Updated webhook flow with this PR: 1. An event happens 2. It is stored as-is and added to a task queue 3. When the task is processed, the event is processed (depending on the webhook type) to make an HTTP request So the only webhook-type dependent logic happens in one place (step 3) which should be much more robust. - the raw event must be stored in the hooktask (until now, the pre-processed body was stored) - to ensure that previous hooktasks are correctly sent, a `payload_version` is added (version 1: the body has already been pre-process / version 2: the body is the raw event) So future webhook additions will only have to deal with creating an http.Request based on the raw event (no need to adjust the code in multiple places, like currently). Moreover since this processing happens when fetching from the task queue, it ensures that the queuing of new events (upon a `git push` for instance) does not get slowed down by a slow webhook. As a concrete example, the PR #19307 for custom webhooks, should be substantially smaller: - no need to change `services/webhook/deliver.go` - minimal change in `services/webhook/webhook.go` (add the new webhook to the map) - no need to change all the individual webhook files (since with this refactor the `*webhook_model.Webhook` is provided as argument) (cherry picked from commit 26653b196bd1d15c532af41f60351596dd4330bd) Conflicts: services/webhook/deliver_test.go trivial context conflict
2024-03-07 23:18:38 +01:00
}
t.ResponseInfo = &webhook_model.HookResponse{
Headers: map[string]string{},
}
// OK We're now ready to attempt to deliver the task - we must double check that it
// has not been delivered in the meantime
updated, err := webhook_model.MarkTaskDelivered(ctx, t)
if err != nil {
log.Error("MarkTaskDelivered[%d]: %v", t.ID, err)
return fmt.Errorf("unable to mark task[%d] delivered in the db: %w", t.ID, err)
}
if !updated {
// This webhook task has already been attempted to be delivered or is in the process of being delivered
log.Trace("Webhook Task[%d] already delivered", t.ID)
return nil
}
// All code from this point will update the hook task
defer func() {
t.Delivered = timeutil.TimeStampNanoNow()
if t.IsSucceed {
log.Trace("Hook delivered: %s", t.UUID)
} else if !w.IsActive {
log.Trace("Hook delivery skipped as webhook is inactive: %s", t.UUID)
} else {
log.Trace("Hook delivery failed: %s", t.UUID)
}
if err := webhook_model.UpdateHookTask(ctx, t); err != nil {
log.Error("UpdateHookTask [%d]: %v", t.ID, err)
}
// Update webhook last delivery status.
if t.IsSucceed {
w.LastStatus = webhook_module.HookStatusSucceed
} else {
w.LastStatus = webhook_module.HookStatusFail
}
if err = webhook_model.UpdateWebhookLastStatus(ctx, w); err != nil {
log.Error("UpdateWebhookLastStatus: %v", err)
return
}
}()
if setting.DisableWebhooks {
return fmt.Errorf("webhook task skipped (webhooks disabled): [%d]", t.ID)
}
if !w.IsActive {
log.Trace("Webhook %s in Webhook Task[%d] is not active", w.URL, t.ID)
return nil
}
resp, err := webhookHTTPClient.Do(req.WithContext(ctx))
if err != nil {
t.ResponseInfo.Body = fmt.Sprintf("Delivery: %v", err)
return fmt.Errorf("unable to deliver webhook task[%d] in %s due to error in http client: %w", t.ID, w.URL, err)
}
defer resp.Body.Close()
// Status code is 20x can be seen as succeed.
t.IsSucceed = resp.StatusCode/100 == 2
t.ResponseInfo.Status = resp.StatusCode
for k, vals := range resp.Header {
t.ResponseInfo.Headers[k] = strings.Join(vals, ",")
}
p, err := io.ReadAll(resp.Body)
if err != nil {
t.ResponseInfo.Body = fmt.Sprintf("read body: %s", err)
return fmt.Errorf("unable to deliver webhook task[%d] in %s as unable to read response body: %w", t.ID, w.URL, err)
}
t.ResponseInfo.Body = string(p)
return nil
}
var (
webhookHTTPClient *http.Client
once sync.Once
hostMatchers []glob.Glob
)
func webhookProxy(allowList *hostmatcher.HostMatchList) func(req *http.Request) (*url.URL, error) {
if setting.Webhook.ProxyURL == "" {
return proxy.Proxy()
}
once.Do(func() {
for _, h := range setting.Webhook.ProxyHosts {
if g, err := glob.Compile(h); err == nil {
hostMatchers = append(hostMatchers, g)
} else {
log.Error("glob.Compile %s failed: %v", h, err)
}
}
})
return func(req *http.Request) (*url.URL, error) {
for _, v := range hostMatchers {
if v.Match(req.URL.Host) {
if !allowList.MatchHostName(req.URL.Host) {
return nil, fmt.Errorf("webhook can only call allowed HTTP servers (check your %s setting), deny '%s'", allowList.SettingKeyHint, req.URL.Host)
}
return http.ProxyURL(setting.Webhook.ProxyURLFixed)(req)
}
}
return http.ProxyFromEnvironment(req)
}
}
// Init starts the hooks delivery thread
func Init() error {
timeout := time.Duration(setting.Webhook.DeliverTimeout) * time.Second
allowedHostListValue := setting.Webhook.AllowedHostList
if allowedHostListValue == "" {
allowedHostListValue = hostmatcher.MatchBuiltinExternal
}
allowedHostMatcher := hostmatcher.ParseHostMatchList("webhook.ALLOWED_HOST_LIST", allowedHostListValue)
webhookHTTPClient = &http.Client{
Timeout: timeout,
Transport: &http.Transport{
TLSClientConfig: &tls.Config{InsecureSkipVerify: setting.Webhook.SkipTLSVerify},
Proxy: webhookProxy(allowedHostMatcher),
DialContext: hostmatcher.NewDialContextWithProxy("webhook", allowedHostMatcher, nil, setting.Webhook.ProxyURLFixed),
},
}
hookQueue = queue.CreateUniqueQueue(graceful.GetManager().ShutdownContext(), "webhook_sender", handler)
if hookQueue == nil {
return fmt.Errorf("unable to create webhook_sender queue")
}
go graceful.GetManager().RunWithCancel(hookQueue)
go graceful.GetManager().RunWithShutdownContext(populateWebhookSendingQueue)
return nil
}
func populateWebhookSendingQueue(ctx context.Context) {
ctx, _, finished := process.GetManager().AddContext(ctx, "Webhook: Populate sending queue")
defer finished()
lowerID := int64(0)
for {
taskIDs, err := webhook_model.FindUndeliveredHookTaskIDs(ctx, lowerID)
if err != nil {
log.Error("Unable to populate webhook queue as FindUndeliveredHookTaskIDs failed: %v", err)
return
}
if len(taskIDs) == 0 {
return
}
lowerID = taskIDs[len(taskIDs)-1]
for _, taskID := range taskIDs {
select {
case <-ctx.Done():
log.Warn("Shutdown before Webhook Sending queue finishing being populated")
return
default:
}
if err := enqueueHookTask(taskID); err != nil {
log.Error("Unable to push HookTask[%d] to the Webhook Sending queue: %v", taskID, err)
}
}
}
}