README
¶
Sentry Hook for Logrus 
Sentry provides both self-hosted and hosted solutions for exception tracking. Both client and server are open source.
Usage
Every sentry application defined on the server gets a different
DSN. In the example below replace
YOUR_DSN with the one created for your application.
import (
"github.com/Sirupsen/logrus"
"github.com/evalphobia/logrus_sentry"
)
func main() {
log := logrus.New()
hook, err := logrus_sentry.NewSentryHook(YOUR_DSN, []logrus.Level{
logrus.PanicLevel,
logrus.FatalLevel,
logrus.ErrorLevel,
})
if err == nil {
log.Hooks.Add(hook)
}
}
If you wish to initialize a SentryHook with tags, you can use the NewWithTagsSentryHook constructor to provide default tags:
tags := map[string]string{
"site": "example.com",
}
levels := []logrus.Level{
logrus.PanicLevel,
logrus.FatalLevel,
logrus.ErrorLevel,
}
hook, err := logrus_sentry.NewWithTagsSentryHook(YOUR_DSN, tags, levels)
If you wish to initialize a SentryHook with an already initialized raven client, you can use
the NewWithClientSentryHook constructor:
import (
"github.com/Sirupsen/logrus"
"github.com/evalphobia/logrus_sentry"
"github.com/getsentry/raven-go"
)
func main() {
log := logrus.New()
client, err := raven.New(YOUR_DSN)
if err != nil {
log.Fatal(err)
}
hook, err := logrus_sentry.NewWithClientSentryHook(client, []logrus.Level{
logrus.PanicLevel,
logrus.FatalLevel,
logrus.ErrorLevel,
})
if err == nil {
log.Hooks.Add(hook)
}
}
hook, err := NewWithClientSentryHook(client, []logrus.Level{
logrus.ErrorLevel,
})
Special fields
Some logrus fields have a special meaning in this hook, and they will be especially processed by Sentry.
| Field key | Description |
|---|---|
event_id |
Each logged event is identified by the event_id, which is hexadecimal string representing a UUID4 value. You can manually specify the identifier of a log event by supplying this field. The event_id string should be in one of the following UUID format: xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx and urn:uuid:xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx) |
user_name |
Name of the user who is in the context of the event |
user_email |
Email of the user who is in the context of the event |
user_id |
ID of the user who is in the context of the event |
user_ip |
IP of the user who is in the context of the event |
server_name |
Also known as hostname, it is the name of the server which is logging the event (hostname.example.com) |
logger |
logger is the part of the application which is logging the event. In go this usually means setting it to the name of the package. |
http_request |
http_request is the in-coming request(*http.Request). The detailed request data are sent to Sentry. |
Timeout
Timeout is the time the sentry hook will wait for a response
from the sentry server.
If this time elapses with no response from the server an error will be returned.
If Timeout is set to 0 the SentryHook will not wait for a reply
and will assume a correct delivery.
The SentryHook has a default timeout of 100 milliseconds when created
with a call to NewSentryHook. This can be changed by assigning a value to the Timeout field:
hook, _ := logrus_sentry.NewSentryHook(...)
hook.Timeout = 20*time.Second
Enabling Stacktraces
By default the hook will not send any stacktraces. However, this can be enabled with:
hook, _ := logrus_sentry.NewSentryHook(...)
hook.StacktraceConfiguration.Enable = true
Subsequent calls to logger.Error and above will create a stacktrace.
Other configuration options are:
StacktraceConfiguration.Levelthe logrus level at which to start capturing stacktraces.StacktraceConfiguration.Skiphow many stack frames to skip before stacktrace starts recording.StacktraceConfiguration.Contextthe number of lines to include around a stack frame for context.StacktraceConfiguration.InAppPrefixesthe prefixes that will be matched against the stack frame to identify it as in_app
Documentation
¶
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type SentryHook ¶
type SentryHook struct {
// Timeout sets the time to wait for a delivery error from the sentry server.
// If this is set to zero the server will not wait for any response and will
// consider the message correctly sent
Timeout time.Duration
StacktraceConfiguration StackTraceConfiguration
// contains filtered or unexported fields
}
SentryHook delivers logs to a sentry server.
func NewSentryHook ¶
func NewSentryHook(DSN string, levels []logrus.Level) (*SentryHook, error)
NewSentryHook creates a hook to be added to an instance of logger and initializes the raven client. This method sets the timeout to 100 milliseconds.
func NewWithClientSentryHook ¶
func NewWithClientSentryHook(client *raven.Client, levels []logrus.Level) (*SentryHook, error)
NewWithClientSentryHook creates a hook using an initialized raven client. This method sets the timeout to 100 milliseconds.
func NewWithTagsSentryHook ¶
func NewWithTagsSentryHook(DSN string, tags map[string]string, levels []logrus.Level) (*SentryHook, error)
NewWithTagsSentryHook creates a hook with tags to be added to an instance of logger and initializes the raven client. This method sets the timeout to 100 milliseconds.
func (*SentryHook) AddExtraFilter ¶ added in v0.2.1
func (hook *SentryHook) AddExtraFilter(name string, fn func(interface{}) interface{})
AddExtraFilter adds a custom filter function.
func (*SentryHook) AddIgnore ¶ added in v0.2.1
func (hook *SentryHook) AddIgnore(name string)
AddIgnore adds field name to ignore.
func (*SentryHook) Fire ¶
func (hook *SentryHook) Fire(entry *logrus.Entry) error
Fire is called when an event should be sent to sentry Special fields that sentry uses to give more information to the server are extracted from entry.Data (if they are found) These fields are: error, logger, server_name and http_request
func (*SentryHook) Levels ¶
func (hook *SentryHook) Levels() []logrus.Level
Levels returns the available logging levels.
type StackTraceConfiguration ¶
type StackTraceConfiguration struct {
// whether stacktraces should be enabled
Enable bool
// the level at which to start capturing stacktraces
Level logrus.Level
// how many stack frames to skip before stacktrace starts recording
Skip int
// the number of lines to include around a stack frame for context
Context int
// the prefixes that will be matched against the stack frame.
// if the stack frame's package matches one of these prefixes
// sentry will identify the stack frame as "in_app"
InAppPrefixes []string
}
StackTraceConfiguration allows for configuring stacktraces
type Stacktracer ¶ added in v0.2.2
type Stacktracer interface {
GetStacktrace() *raven.Stacktrace
}
Source Files