Defer Panic Client Lib.
-
Panic Handling - Let the client catch and log any panic you get to your own dashboard.
-
Error Handling - Handle your errrors - but log them.
-
HTTP latency - DeferClient can log the latencies of all your hard hit http requests.
-
HTTP RPM - Aggregate all the response codes for all your HTTP requests over time.
-
Micro Services Tracing - DeferClient can log queries to other services so you know exactly what service is slowing down the initial request!
-
Database latency - Get notified of slow database queries in your go app.
-
Metrics - See:
- go routines
- memory
- usage
- gc runs && pause times
- cgo
- file descriptors over time
and more automatically in your own dashboard.
-
Expvar - Got expvars already in your app - we support it out of the box.
-
Custom K/V - Got something we don't support? You can log your own k/v metrics just as easily. Use this client.
Translations:
go get github.com/deferpanic/deferclient/deferstats
api key
Get an API KEY via your shell or signup manually here:
curl https://api.deferpanic.com/v1.13/users/create \
-X POST \
-d "email=test@test.com" \
-d "password=password"
Here we have 3 examples:
- log a fast request
- log a slow request
- log a panic
package main
import (
"fmt"
"github.com/deferpanic/deferclient/deferstats"
"net/http"
"time"
)
func panicHandler(w http.ResponseWriter, r *http.Request) {
panic("there is no need to panic")
}
func fastHandler(w http.ResponseWriter, r *http.Request) {
fmt.Fprintf(w, "this request is fast")
}
func slowHandler(w http.ResponseWriter, r *http.Request) {
time.Sleep(2 * time.Second)
fmt.Fprintf(w, "this request is slow")
}
func main() {
dps := deferstats.NewClient("v00L0K6CdKjE4QwX5DL1iiODxovAHUfo")
go dps.CaptureStats()
http.HandleFunc("/fast", dps.HTTPHandlerFunc(fastHandler))
http.HandleFunc("/slow", dps.HTTPHandlerFunc(slowHandler))
http.HandleFunc("/panic", dps.HTTPHandlerFunc(panicHandler))
http.ListenAndServe(":3000", nil)
}You can see that our slow request triggers our latency threshold and winds up in our dashboard while the panic is also caught and winds up in the dashboard.
The default latency threshold is 500ms.
The client works perfectly fine in non-HTTP applications:
If you want to log your errors use this method. deferstats.Wrap will log the bactrace and the error and ship it up to deferpanic immediately.
package main
import (
"errors"
"fmt"
"github.com/deferpanic/deferclient/deferstats"
"time"
)
var (
dps *deferstats.Client
)
func errorTest() {
err := errors.New("danger will robinson!")
if err != nil {
dps.Wrap(err)
fmt.Println(err)
}
}
func panicTest() {
defer dps.Persist()
panic("there is no need to panic")
}
func main() {
dps = deferstats.NewClient("v00L0K6CdKjE4QwX5DL1iiODxovAHUfo")
errorTest()
panicTest()
time.Sleep(time.Second * 20)
}package main
import (
"database/sql"
"github.com/deferpanic/deferclient/deferstats"
_ "github.com/lib/pq"
"log"
"time"
)
func main() {
dps := deferstats.NewClient("v00L0K6CdKjE4QwX5DL1iiODxovAHUfo")
_db, err := sql.Open("postgres", "dbname=dptest sslmode=disable")
db := deferstats.NewDB(_db)
go dps.CaptureStats()
var id int
var sleep string
err = db.QueryRow("select 1 as num, pg_sleep(0.25)").Scan(&id, &sleep)
if err != nil {
log.Println("oh no!")
}
err = db.QueryRow("select 1 as num, pg_sleep(2)").Scan(&id, &sleep)
if err != nil {
log.Println("oh no!")
}
time.Sleep(120 * time.Second)
}
We have additional database ORM wrappers:
package main
import (
"expvar"
"fmt"
"github.com/deferpanic/deferclient/deferstats"
"github.com/deferpanic/deferclient/expvars"
"io"
"net/http"
)
var (
counter int64
inthits = expvar.NewInt("inthits")
floathints = expvar.NewFloat("floathints")
stringhints = expvar.NewString("stringhits")
maphints = expvar.NewMap("maphints")
)
func init() {
maphints.Set("int", inthits)
maphints.Set("float", floathints)
maphints.Set("string", stringhints)
}
func hit(w http.ResponseWriter, r *http.Request) {
counter++
inthits.Set(counter)
floathints.Set(float64(counter * 2))
stringhints.Set(fmt.Sprintf("%v", counter*3))
io.WriteString(w, fmt.Sprintf("%v", counter))
}
func main() {
dps := deferstats.NewClient("v00L0K6CdKjE4QwX5DL1iiODxovAHUfo")
dps.GrabExpvar = true
dps.GetExpvar = expvars.GetByDo
go dps.CaptureStats()
http.HandleFunc("/hit", hit)
http.ListenAndServe(":8000", nil)
}
Got a micro-services/SOA architecture? Now you can trace your queries throughout and tie them back to the initial query that started it all!
We set the HTTP header X-dpparentspanid in each request to trace requests throughout your architecture.
Example
User Facing Service
package main
import (
"fmt"
"github.com/deferpanic/deferclient/deferstats"
"io/ioutil"
"net/http"
)
func handler(w http.ResponseWriter, r *http.Request) {
// just pass your spanId w/each request
client := &http.Client{}
request, err := http.NewRequest("POST", "http://127.0.0.1:7070/internal", nil)
if err != nil {
fmt.Println(err)
}
request.Header.Add("X-dpparentspanid", deferstats.GetSpanIdString(w))
resp, err := client.Do(request)
if err != nil {
fmt.Println(err)
}
defer resp.Body.Close()
body, err := ioutil.ReadAll(resp.Body)
fmt.Fprintf(w, string(body))
}
func main() {
dfs := deferstats.NewClient("v00L0K6CdKjE4QwX5DL1iiODxovAHUfo")
go dfs.CaptureStats()
http.HandleFunc("/", dfs.HTTPHandlerFunc(handler))
http.ListenAndServe(":9090", nil)
}Slow Internal API
package main
import (
"encoding/json"
"github.com/deferpanic/deferclient/deferstats"
"net/http"
"time"
)
type blah struct {
Stuff string
}
func handler(w http.ResponseWriter, r *http.Request) {
time.Sleep(510 * time.Millisecond)
stuff := blah{
Stuff: "some reply",
}
js, err := json.Marshal(stuff)
if err != nil {
http.Error(w, err.Error(), http.StatusInternalServerError)
return
}
w.Header().Set("Content-Type", "application/json")
w.Write(js)
}
func main() {
dfs := deferstats.NewClient("v00L0K6CdKjE4QwX5DL1iiODxovAHUfo")
go dfs.CaptureStats()
http.HandleFunc("/internal", dfs.HTTPHandlerFunc(handler))
http.ListenAndServe(":7070", nil)
}Notice that when you wrap your http handlers w/the deferstats httphandler we instantly tie the front facing service to the slow internal one.
Want to monitor both staging and production? By default the environment is set to 'production' but you can us a different environment just by setting the environment variable to wahtever you wish and all data will be tagged this way. Then you can change it from your settings in your dashboard.
package main
import (
"github.com/deferpanic/deferclient/deferstats"
"time"
)
func main() {
dfs := deferstats.NewClient("v00L0K6CdKjE4QwX5DL1iiODxovAHUfo")
dfs.Setenvironment("some-other-environment")
go dfs.CaptureStats()
time.Sleep(120 * time.Second)
}Many deferPanic users are using micro-servіces/SOA and sometimes you want to see activity from one application versus all of them.
You can set this via the AppGroup variable and then toggle it from your settings in your dashboard.
package main
import (
"github.com/deferpanic/deferclient/deferstats"
"time"
)
func main() {
dfs := deferstats.NewClient("v00L0K6CdKjE4QwX5DL1iiODxovAHUfo")
dfs.SetappGroup("queueMaster3000")
go dfs.CaptureStats()
time.Sleep(120 * time.Second)
}Pretty much everyone needs to disable the client in a dev/test mode otherwise you are going to be overwhelmed with notifications.
You can disable it simply by using the setter function SetnoPost.
package main
import (
"github.com/deferpanic/deferclient/deferstats"
"time"
)
func main() {
dfs := deferstats.NewClient("v00L0K6CdKjE4QwX5DL1iiODxovAHUfo")
dfs.SetnoPost(true)
go dfs.CaptureStats()
time.Sleep(120 * time.Second)
}There are currently no dependencies so this should work out of the box for container runtimes like docker && rocket.
See https://godoc.org/github.com/deferpanic/deferclient for documentation.
Defer Panic Client
Q: Does this support websockets yet?
A: No but we are open to discussing what this would look like.
Q: Do you all support Google App Engine?
A: Not at this time but we do support Google Compute Engine.