cloud-test

As the summer heat descends on the Northern Hemisphere we thought we’d release our newest App Engine version with some changes that are sure to keep you playing around in the cool, air-conditioned indoors (hey, you don’t want your computer to overheat, right?).Production Changes

Adjustable Scheduler Parameters - As we previously discussed, we are introducing two scheduler knobs (okay, they actually look like sliders) that will allow you to control some of the parameters that influence how many Instances run your application. Today you will be able to set the minimum pending latency and maximum number of idle instances for your application.

Datastore Changes

Advanced Query Planning - We are removing the need for exploding indexes and reducing the custom index requirements for many queries. The SDK will suggest better indexes in several cases and an upcoming article will describe what further optimizations are possible.

Namespaced Datastore Stats - Now, in addition to getting overall datastore stats, we are providing a new option to query datastore stats per namespace.

Task Queue Changes

New Task Queue details page - We’ve revamped the Task Queue details page in the Administration Console to provide more information about the tasks being run. You can now see the headers included in the enqueued task, the payload, and information from previous task runs.

1MB Pull Task Size - It’s our belief that there is only one way for size limits to go - and that’s up! So with this release we’ve increased the size for pull tasks to 1MB.

Pull queue lease modification - We’ve introduced a new method for Pull Queues that allows you to extend the lease on existing tasks if the initial lease on the task was insufficient.

Lastly, we have some exciting news related to the experimental Go runtime. While it still remains experimental, starting with 1.5.2, all HRD apps will have access to the Go runtime in production.As always, there are also some small features and bug fixes, the full list of which can be found in our release notes (Python, Java). We look forward to your feedback and questions in our forum.Posted by The App Engine Team

Google

Postmortem: Java App Engine outage, July 14, 2011

Wednesday, July 20, 2011

Summary

Last week, we posted about a limited outage on July 14, 2011. Now that our internal postmortem is complete, we thought you would also like to get more detail about what went wrong and what we are going to do to ensure this doesn't happen again.

Root Cause and Analysis

The main lesson learned is to improve our live traffic testing as a relatively minor bug triggered a corner case for some of our customers. The bug was in a new release of the infrastructure in the App Engine Java execution environment. During development, testing, and qualification, this bug was essentially hidden from view because it only manifested itself under specific load patterns. During the outage, requests to affected applications would fail with errors when traffic was routed to affected instances. Application logs would have shown affected instances experienced high latency, error rates, or were not reachable from the Internet. This could have been caught by letting the live traffic testing run longer.

In order for live traffic testing to work properly, we need to improve our monitoring as well. In this case, having more points from which to do black box monitoring would have helped immensely. We are currently working on much broader monitoring for App Engine and will be integrating more extensive black box testing in upcoming quarters.

Once again, we’d like to point out that we could have done a much better job of communicating issues to all of you. While we strive to strike a balance between letting you know about major issues and not bothering you about the day-to-day operations; we clearly should have communicated this incident to you sooner. Rest assured you’ll be better informed of issues in the future.

Timeline

July 14, 2011 - 11:30 AM US/Pacific - The new Java execution environment is released to production.

July 14, 2011 - 5:00-6:00 PM US/Pacific - The previously scheduled Master/Slave read-only maintenance period occurred.

July 14, 2011 - 8:00-9:30 PM US/Pacific - Monitoring shows error rates and latency for Java applications using the Master/Slave datastore are slowly increasing across the entire system. Investigation reveals that the new Java execution environment is malfunctioning.

July 14, 2011 - 9:30 PM US/Pacific - Rollback of the Java execution environment to the previous version begins. Latency and error rates begin to fall.

July 14, 2011 - 11:30 PM US/Pacific - Rollback of the Java execution environment to the previous version completes. Java Master/Slave applications are functioning normally.

Remediation

Faster notification on our status site and downtime-notify mailing list

More live traffic stress tests for new releases

Better black box monitoring to detect small impacts more quickly

[Edit] Clarification: no HR datastore apps were affected. Overall, the outage resulted in a 1.9% error rate, affecting approximately 0.005% of all App Engine traffic at peak.

Posted by Wesley Chun, Google App Engine Team

Google

Java App Engine outage, July 14, 2011

Friday, July 15, 2011

On July 14, 2011, beginning at 7 PM US/Pacific time (PDT/GMT-7), a subset of Java App Engine applications were affected by a service outage, which gradually increased in magnitude over time. At 9:30 PM US/Pacific, repair work commenced which began to reduce the effect of the outage; by 11:30 PM US/Pacific, the repair work had completed, restoring normal service to all Java App Engine applications.

During this period, affected applications would have experienced high latency and error rates. This outage occurred shortly after a scheduled maintenance period; however, the outage was not related to the maintenance work.

Overall reliability, quick return to service, and fast, accurate communication to our customers are some of the core goals of Google App Engine's service offering. While we restored service relatively quickly, it's clear to us that we fell short in prompt communication of status updates. We apologize for this, and we'll look at our procedures to improve our performance in this area.

In the meantime, we have a preliminary understanding of the outage, and we are continuing our investigation to insure that we have fully repaired the root cause. We will publish a detailed postmortem once we have concluded our research. Thanks again for your patience and understanding.

[Edit] Clarification: no HR datastore apps were affected. Overall, the outage resulted in a 1.9% error rate, affecting approximately 0.005% of all App Engine traffic at peak.

Posted by Wesley Chun, Google App Engine team

Google

Wanted: App Engineers

Wednesday, July 13, 2011

You may have heard the news that 2011 is Google’s biggest hiring year yet. And the App Engine team is looking for a few great Software Engineers to join us in San Francisco to code, collaborate on innovative ideas for platform computing, and get burritos with us every Friday.

The App Engine team in San Francisco.

We think we have some of the best and most enthusiastic developers out there-- so we thought we’d ask you to come work with us. The App Engine team needs talented developers to help build the platform by developing the features on our road map and issue tracker, as well as that amazing feature idea you have that will revolutionize App Engine for our users.If you are a backend engineer who is experienced with C++ or Java, lives and breathes distributed systems, and is obsessed with scalability, please visit our jobs website to submit your resume along with a cover letter describing your experience with App Engine. We’ll be in touch if we think your experience matches our open positions.Thanks for continuing to help make App Engine a great platform!Posted by The App Engine Team

Google

Highlights from our Google I/O Presentations

Wednesday, June 22, 2011

Now that things have settled down from last month’s Google I/O, we thought it would be a good time to highlight some of the technical talks given by the App Engine team. At this year’s I/O, we emphasized some of the big themes we’ve been focusing on this past year: availability and productionization of our platform, removing limits while maintaining performance at scale, and developing new APIs.Availability and Productionization of our PlatformSince App Engine automatically handles all of the sharding and distribution of your application across datacenters, we thought it would be important to talk about the strategies we use to run and maintain our platform. These talks also give you an opportunity to meet some of the engineers who carry pagers so you don’t have to.

More 9s Please: Under the Covers of the High Replication Datastore - We were excited to announce the HR Datastore in January, since it provides a high level of availability, even during unplanned outages, and even if an entire datacenter goes offline. HRD uses the paxos algorithm to distribute data across multiple datacenters. Alfred Fuller and Matt Wilder talk about the high level distribution strategy used and how this helps keep your app serving.

Life in App Engine Production - Let’s face it - everyone has some great war stories about running live applications. And the stories of Google’s SRE team probably rank with the best of them. Michael Handler and Alan Green talk about what goes right and what can go wrong running in Production.

Removing Limits while maintaining performance at scaleHow can you design your application so you can transition from serving 1 user to 1 million users with the minimum amount of work? How do you make sure your user experience stays consistent for those users?

App Engine Backends - At I/O this year we launched App Engine Backends, which allow you to write long running processes on App Engine. Crunch numbers, process data, and build large in-memory models! Justin Haugh and Greg Darke give an overview of this new feature and cover best practices.

Scaling App Engine Applications - We always get lots of question on what it means and how it works to scale an app using App Engine (hint: low latency is the key!). Guido van Rossum and Justin Haugh explain how instances work and cover best practices for scaling your application.

Developing New APIs

Full Text Search - It took 8 days after App Engine’s initial launch for this feature request to be filed - and now work on the Full Text Search API is nearing completion. Bo Majewski and Ged Ellis give a preview of the upcoming App Engine Full Text Search API.

App Engine MapReduce- So you just want to analyze 5TB of data? You’ll probably need MapReduce to do so! Mike Aizatsky talks about the progress being made on the App Engine MapReduce framework and how you can use it to generate reports for your data.

Large-scale Data Analysis using the App Engine Pipeline API - Once you’ve written one MapReduce, you’ll probably want to write more, and then you’ll need a way to manage your application’s analysis pipeline. In this talk, Brett Slatkin discusses a lightweight API for managing MapReduce workflows.

Our team always enjoys participating in Google I/O because our engineers get to share their vision and discuss their work with a large audience. If you are interested in what some of our customers had to say at I/O, check out the Google Enterprise Blog. Of course, with App Engine featured in 23 talks for Google I/O, we couldn’t cover them all here. See the complete session list for I/O and discover all the talks on App Engine and more!Posted by The App Engine Team

Google

App Engine 1.5.1 Release

Monday, June 20, 2011

It’s been about a month since Google I/O 2011, so we thought it was time for another release. This month we are releasing ProtoRPC as an official Python API, making the High Replication Datastore (HRD) available in the SDK to help developers better understand the consistency model, adding presence to the Channel API, plus a few more goodies. Check them out below:Serving Changes

Geolocation Header: App Engine will now include a header with every client request that represents a best-effort attempt to identify the country from which the request originated. This header, “X-AppEngine-country,” can allow you to customize content based on the origin of the user. We hope this means no more “Select your location” drop downs for users!

Updated Java and Python APIs

Channel API (Presence Support): You can now configure an inbound service for your application to detect user presence for the Channel API. This lets you take an action when a user connects or disconnects from the Channel.

Images API (WebP Support): The Images API now supports WebP images. If you haven’t checked it out, WebP is a new image format that Google open sourced earlier this year which provides lossy compression for images and can be up to 39% smaller than JPEG images of a similar quality!

New Python APIs

ProtoRPC: ProtoRPC is an open source framework for creating a well defined and easy to use web-based RPC service. While ProtoRPC is similar to using Google Protocol Buffers to define services, the goal of ProtoRPC is to make it easy for developers to start defining web based services, while still allowing those services to scale and evolve over time.

Datastore

High Replication in the SDK: Since releasing the High Replication Datastore, we’ve wanted to provide tools that help developers understand and test the new consistency model while developing their applications. The 1.5.1 SDKs for Java and Python can now emulate the HRD consistency model. This means that now, by setting the appropriate SDK config options, queries across entity groups will occasionally return results that don’t reflect the most recent data written. This should allow you to develop your application to be more resilient to this consistency model.

As with all releases, 1.5.1 contains some additional small features and many bugfixes, the full list of which is available in our release notes for Python and Java. Hopefully you’ll find this release useful. Either way feel free let us know what you think by posting in the App Engine group.Thanks!The App Engine Team