Google Cloud Platform Blog
Diagnose Service Performance Bottlenecks with Google Cloud Trace Beta
Thursday, January 8, 2015
Here at Google, we understand the importance of having applications run at optimal speed. Awesome performance of your application is critical for
end user satisfaction and retention
. User expectations for application performance are already high and applications with poor performance
risk losing users
. For example,
25% of users abandon a web page if its load time is more than four seconds
and
86% of users delete an application after poor performance
. Ask any developer who has experienced the stress of diagnosing performance issues in production and you will find that it is extremely difficult to isolate the root cause of poor performance when it happens. This is especially true when the sluggish behavior is only seen by a small fraction of your users.
We
introduced
Google Cloud Trace at Google I/O 2014 and gave an in-depth
talk and demo
at Google Cloud Platform Live back in November. Today, we’re announcing the beta release of Google Cloud Trace which is now available to all Google Cloud Platform customers.
With Google Cloud Trace, you can diagnose performance issues in your production application by quickly finding the traces for slow requests and viewing a detailed report of where time is spent in your application while processing these requests. Its trace analysis feature allows you to see the latency distribution for your application, and find the painfully slow requests that may be affecting only a small number of your users. You can also use the trace analysis feature to check if the performance of a new release is better than the previous release.
If you look at the screenshots below, you can get a feel for the type of insight Cloud Trace provides. Figure 1 shows the breakdown of a single request to an application, and where time was spent in processing the request.
Figure 1 - Breakdown of a single request
Cloud Trace can analyze a set of requests to show their latency distribution, percentile latency values with sample traces, and the remote procedure calls that are significant latency bottlenecks.
Figure 2 - Request analysis
Cloud Trace also allows you to compare the latency profiles for an application’s requests from before and after a change is made, such as rolling out a patch. Figure 3 shows the comparison between the before and after requests.
Figure 3 - Comparing requests
To get started with using Cloud Trace for your project, visit your project’s home page in the
Google Developers Console
, select
Monitoring > Traces
in the navigation panel, click on the
Settings
tab, and set
Traces
to
On
. Cloud Trace has very little overhead, so you can safely leave traces enabled for your application without worrying about performance impact. Once traces are enabled for your application, Cloud Trace will start tracing requests received by your application, and continue to do so as you deploy new versions of your application.
Take a look at our
Google Cloud Platform Live talk
, check out our
documentation
, and
try it yourself
-- there’s no set up required. All you need is an application running on App Engine or a Managed VM. We look forward to receiving your
direct feedback
.
-Posted by Pratul Dublish, Product Manager
No comments :
Post a Comment
Don't Miss Next '17
Use promo code NEXT1720 to save $300 off general admission
REGISTER NOW
Free Trial
GCP Blogs
Big Data & Machine Learning
Kubernetes
GCP Japan Blog
Labels
Announcements
56
Big Data & Machine Learning
91
Compute
156
Containers & Kubernetes
36
CRE
7
Customers
90
Developer Tools & Insights
80
Events
34
Infrastructure
24
Management Tools
39
Networking
18
Open Source
105
Partners
63
Pricing
24
Security & Identity
23
Solutions
16
Stackdriver
19
Storage & Databases
111
Weekly Roundups
16
Archive
2017
Feb
Jan
2016
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2015
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2014
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2013
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2012
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2011
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2010
Dec
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2009
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
2008
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Feed
Subscribe by email
Technical questions? Check us out on
Stack Overflow
.
Subscribe to
our monthly newsletter
.
Google
on
Follow @googlecloud
Follow
Follow
No comments :
Post a Comment