Ton of minor fixes all pulled in. More info in the latest release for 1.2.0
This commit is contained in:
+190
-16
@@ -11,6 +11,7 @@ import (
|
||||
"io"
|
||||
"io/ioutil"
|
||||
"log"
|
||||
"math/rand"
|
||||
"net/http"
|
||||
"net/url"
|
||||
"os"
|
||||
@@ -251,16 +252,42 @@ func handleGetWorkflowqueue(resp http.ResponseWriter, request *http.Request) {
|
||||
return
|
||||
}
|
||||
|
||||
id := request.Header.Get("Org-Id")
|
||||
if len(id) == 0 {
|
||||
log.Printf("[INFO] No org-id header set")
|
||||
// This is really the environment's name - NOT org-id
|
||||
orgId := request.Header.Get("Org-Id")
|
||||
if len(orgId) == 0 {
|
||||
log.Printf("[AUDIT] No org-id header set")
|
||||
resp.WriteHeader(401)
|
||||
resp.Write([]byte(fmt.Sprintf(`{"success": false, "reason": "Specify the org-id header."}`)))
|
||||
return
|
||||
}
|
||||
|
||||
ctx := context.Background()
|
||||
env, err := shuffle.GetEnvironment(ctx, id, "")
|
||||
environment := request.Header.Get("org")
|
||||
if len(environment) == 0 {
|
||||
log.Printf("[AUDIT] No 'org' header set (get workflow queue). Required for cloud.")
|
||||
/*
|
||||
resp.WriteHeader(403)
|
||||
resp.Write([]byte(fmt.Sprintf(`{"success": false, "reason": "Specify the org header. This can be done by setting the 'ORG' environment variable for Orborus to your Org ID in Shuffle"}`)))
|
||||
return
|
||||
*/
|
||||
}
|
||||
|
||||
orborusLabel := request.Header.Get("x-orborus-label")
|
||||
|
||||
// This section is cloud custom for now
|
||||
auth := request.Header.Get("Authorization")
|
||||
if len(auth) == 0 {
|
||||
log.Printf("[AUDIT] No Authorization header set. Required for cloud. Env: %s, org: %s", orgId, environment)
|
||||
/*
|
||||
resp.WriteHeader(401)
|
||||
resp.Write([]byte(fmt.Sprintf(`{"success": false, "reason": "Specify the auth header (only applicable for cloud for now)."}`)))
|
||||
return
|
||||
*/
|
||||
}
|
||||
|
||||
//log.Printf("[AUDIT] Get workflow queue for org %s, env %s, orborus label %s", orgId, environment, orborusLabel)
|
||||
|
||||
ctx := shuffle.GetContext(request)
|
||||
env, err := shuffle.GetEnvironment(ctx, orgId, "")
|
||||
timeNow := time.Now().Unix()
|
||||
if err == nil && len(env.Id) > 0 && len(env.Name) > 0 {
|
||||
if time.Now().Unix() > env.Edited+60 {
|
||||
@@ -273,7 +300,154 @@ func handleGetWorkflowqueue(resp http.ResponseWriter, request *http.Request) {
|
||||
}
|
||||
}
|
||||
|
||||
executionRequests, err := shuffle.GetWorkflowQueue(ctx, id, 100)
|
||||
//log.Printf("Found env: %#v", env)
|
||||
if len(env.OrgId) > 0 {
|
||||
environment = env.OrgId
|
||||
}
|
||||
|
||||
if request.Method == "POST" {
|
||||
if rand.Intn(1) == 0 {
|
||||
// Parse out body
|
||||
body, err := ioutil.ReadAll(request.Body)
|
||||
if err == nil {
|
||||
|
||||
// Parse out CPU, memory and disk.
|
||||
|
||||
var envData shuffle.OrborusStats
|
||||
err = json.Unmarshal(body, &envData)
|
||||
if err == nil && !envData.Swarm && !envData.Kubernetes && (envData.CPU > 0 || envData.Memory > 0 || envData.Disk > 0) {
|
||||
|
||||
// Set the input in memory
|
||||
envData.OrgId = orgId
|
||||
envData.Environment = environment
|
||||
envData.OrborusLabel = orborusLabel
|
||||
envData.Timestamp = time.Now().Unix()
|
||||
|
||||
if envData.CPU > 0 && envData.MaxCPU > 0 {
|
||||
envData.CPUPercent = float64(envData.CPU) / float64(envData.MaxCPU)
|
||||
}
|
||||
|
||||
if envData.Memory > 0 && envData.MaxMemory > 0 {
|
||||
envData.MemoryPercent = float64(envData.Memory) / float64(envData.MaxMemory)
|
||||
}
|
||||
|
||||
// Check if CPU percent constantly has stayed above X% for the last Y requests
|
||||
percentageCheck := 90
|
||||
concurrentChecks := 0
|
||||
|
||||
//if int(envData.CPUPercent) > percentageCheck {
|
||||
// Get cached data
|
||||
percentages := []float64{}
|
||||
cacheKey := fmt.Sprintf("%s_%s_percent", orgId, strings.ToLower(environment))
|
||||
|
||||
// Marshal float list into []byte
|
||||
cacheData := []byte{}
|
||||
cache, err := shuffle.GetCache(ctx, cacheKey)
|
||||
if err == nil {
|
||||
// Unmarshal into percentages
|
||||
cacheData := []byte(cache.([]uint8))
|
||||
err = json.Unmarshal(cacheData, &percentages)
|
||||
if err != nil {
|
||||
log.Printf("[INFO] error in cache unmarshal for percentages: %s", err)
|
||||
}
|
||||
|
||||
if len(percentages) > concurrentChecks {
|
||||
percentages = percentages[:concurrentChecks]
|
||||
}
|
||||
|
||||
percentages = append(percentages, envData.CPUPercent)
|
||||
if len(percentages) > concurrentChecks {
|
||||
//log.Printf("[INFO] Checking percentages: %v", percentages)
|
||||
|
||||
// percentageCheck := 1
|
||||
sendAlert := true
|
||||
for _, p := range percentages {
|
||||
if int(p) < percentageCheck {
|
||||
//log.Printf("[AUDIT] CPU percent is below %d: %d", percentageCheck, int(p))
|
||||
sendAlert = false
|
||||
break
|
||||
}
|
||||
}
|
||||
|
||||
if sendAlert {
|
||||
log.Printf("[INFO] CPU percent has been above %d percent for the last 5 requests. Sending alert. Env: %s, org: %s", percentageCheck, environment, orgId)
|
||||
|
||||
// Set notification + alert for organization
|
||||
err = shuffle.CreateOrgNotification(
|
||||
ctx,
|
||||
fmt.Sprintf("CPU percent has been above %d percent", percentageCheck),
|
||||
fmt.Sprintf("A environment %s has been using more than %d\\% CPU for the last 5 requests.", environment, percentageCheck),
|
||||
fmt.Sprintf("/admin?tab=environments"),
|
||||
environment,
|
||||
true,
|
||||
)
|
||||
|
||||
if err != nil {
|
||||
log.Printf("[ERROR] error creating notification: %s", err)
|
||||
}
|
||||
|
||||
org, err := shuffle.GetOrg(ctx, environment)
|
||||
if err == nil {
|
||||
foundRecommendation := false
|
||||
for _, recommendation := range org.Priorities {
|
||||
if strings.Contains(recommendation.Name, "CPU") {
|
||||
foundRecommendation = true
|
||||
break
|
||||
}
|
||||
}
|
||||
|
||||
if !foundRecommendation {
|
||||
// Add to start of org.Priorities
|
||||
org.Priorities = append(org.Priorities, shuffle.Priority{
|
||||
Name: fmt.Sprintf("High CPU in environment %s", orgId),
|
||||
Description: fmt.Sprintf("The environment %s has been using more than %d percent CPU.", orgId, percentageCheck),
|
||||
Type: "scale",
|
||||
Active: true,
|
||||
URL: fmt.Sprintf("/admin?tab=environments"),
|
||||
})
|
||||
|
||||
//Make last item the first item
|
||||
org.Priorities = append([]shuffle.Priority{org.Priorities[len(org.Priorities)-1]}, org.Priorities[:len(org.Priorities)-1]...)
|
||||
err = shuffle.SetOrg(ctx, *org, org.Id)
|
||||
if err != nil {
|
||||
log.Printf("[ERROR] Problem setting org: %s", err)
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if len(percentages) > 1 {
|
||||
percentages = percentages[1:]
|
||||
}
|
||||
}
|
||||
|
||||
// Marshal float list into []byte
|
||||
} else {
|
||||
//log.Printf("[ERROR] Failed getting cache: %s", err)
|
||||
percentages = append(percentages, envData.CPUPercent)
|
||||
}
|
||||
|
||||
if len(percentages) > 0 {
|
||||
//log.Printf("[DEBUG] Setting cache for %s: %#v", cacheKey, percentages)
|
||||
cacheData, err = json.Marshal(percentages)
|
||||
if err != nil {
|
||||
log.Printf("[INFO] error in cache marshal: %s", err)
|
||||
}
|
||||
|
||||
// Add the new data
|
||||
go shuffle.SetCache(ctx, cacheKey, cacheData, 5)
|
||||
}
|
||||
}
|
||||
|
||||
//log.Printf("CPU percent: %f", envData.CPUPercent)
|
||||
//log.Printf("Memory percent: %f", envData.MemoryPercent*100)
|
||||
|
||||
go shuffle.SetenvStats(ctx, envData)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
executionRequests, err := shuffle.GetWorkflowQueue(ctx, orgId, 100)
|
||||
if err != nil {
|
||||
// Skipping as this comes up over and over
|
||||
//log.Printf("(2) Failed reading body for workflowqueue: %s", err)
|
||||
@@ -290,21 +464,21 @@ func handleGetWorkflowqueue(resp http.ResponseWriter, request *http.Request) {
|
||||
|
||||
// Try again :)
|
||||
if len(env.Id) == 0 && len(env.Name) == 0 {
|
||||
orgId := ""
|
||||
foundId := ""
|
||||
for _, requestData := range executionRequests.Data {
|
||||
execution, err := shuffle.GetWorkflowExecution(ctx, requestData.ExecutionId)
|
||||
if err == nil {
|
||||
if len(execution.ExecutionOrg) > 0 {
|
||||
orgId = execution.ExecutionOrg
|
||||
foundId = execution.ExecutionOrg
|
||||
break
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if len(orgId) > 0 {
|
||||
env, err := shuffle.GetEnvironment(ctx, id, orgId)
|
||||
env, err := shuffle.GetEnvironment(ctx, orgId, foundId)
|
||||
if err != nil {
|
||||
log.Printf("[WARNING] No env found matching %s - continuing without updating orborus anyway: %s", id, err)
|
||||
log.Printf("[WARNING] No env found matching %s - continuing without updating orborus anyway: %s", orgId, err)
|
||||
//resp.WriteHeader(401)
|
||||
//resp.Write([]byte(fmt.Sprintf(`{"success": false, "reason": "No env found matching %s"}`, id)))
|
||||
//return
|
||||
@@ -361,9 +535,9 @@ func handleGetStreamResults(resp http.ResponseWriter, request *http.Request) {
|
||||
err = json.Unmarshal(body, &actionResult)
|
||||
if err != nil {
|
||||
log.Printf("[WARNING] Failed ActionResult unmarshaling (stream result): %s", err)
|
||||
resp.WriteHeader(401)
|
||||
resp.Write([]byte(fmt.Sprintf(`{"success": false, "reason": "%s"}`, err)))
|
||||
return
|
||||
//resp.WriteHeader(401)
|
||||
//resp.Write([]byte(fmt.Sprintf(`{"success": false, "reason": "%s"}`, err)))
|
||||
//return
|
||||
}
|
||||
|
||||
ctx := context.Background()
|
||||
@@ -476,9 +650,9 @@ func handleWorkflowQueue(resp http.ResponseWriter, request *http.Request) {
|
||||
err = json.Unmarshal(body, &actionResult)
|
||||
if err != nil {
|
||||
log.Printf("[WARNING] Failed ActionResult unmarshaling (queue): %s", err)
|
||||
resp.WriteHeader(401)
|
||||
resp.Write([]byte(fmt.Sprintf(`{"success": false, "reason": "%s"}`, err)))
|
||||
return
|
||||
//resp.WriteHeader(401)
|
||||
//resp.Write([]byte(fmt.Sprintf(`{"success": false, "reason": "%s"}`, err)))
|
||||
//return
|
||||
}
|
||||
|
||||
//log.Printf("Received action: %#v", actionResult)
|
||||
|
||||
Reference in New Issue
Block a user