5 ms·
I wish there was something better than Elastic for indexing large volume of text.
by StreamBright 6y ago
I wish there was something better than Elastic for indexing large volume of text.
- k__ 6y agoYes. AWS should build a real serverless alternative to it or buy Algolia or something...
- StreamBright 6y agoYes, I would be much happier that. I do not even understand why to do this whole dance with Elastic.
- lnsp 6y agoA lot of people have adopted the Elastic stack over the years, I guess Amazon just wanted their share.
- uncledave 6y agoI think most people just wanted Amazon to take the administrative overhead away from them. Amazon saw that gap.
- k__ 6y agoSure thing. That's the reason for many services they created. But they usually tend to build their own alternative that integrates better into their eco-system.
- brodouevencode 6y agoThey do. They've done similar things with Redis and MariaDB, both of whom also offer commercial solutions. The SAs all insist that "no it's a partnership". I don't fault them they're parroting the marketing line. With the new Grafana stuff that seems more like a real partnership because you have to purchase the enterprise features and support from the marketplace. What they've done to Redis and partiularly Elastic is pretty shameful. The Elasticsearch Service has problems: you can't join the cluster like you can with the normal version, permissions are janky at best, it's slow, and it's expensive. For those reasons the projects my team used it on opted to roll our own elasticsearch cluster which proved a better solution long term beyond the initial annoyances. I say that to say it's not a one to one product, which will probably be their defense.
- jokethrowaway 6y agoSame. But there is nothing preventing doing that and running a fork of Elastic, it's cheap enough. I would do the same if I were AWS.
- lovelearning 6y agoKendra is their Algolia-like service. https://docs.aws.amazon.com/kendra/latest/dg/in-adding-documents.html https://docs.aws.amazon.com/kendra/latest/dg/in-adding-docum...
- k__ 6y agoBut not serverless. If you create an index, you have to pay at least 2,50€ per hour it runs. All pricing examples were multiple thousands of Euros per month.
- buro9 6y agohttps://grafana.com/oss/loki/ https://grafana.com/oss/loki/
- uncledave 6y agoThat’s absolutely not the right tool for indexing large volumes of text. It’s a time series database with very rudimentary labelling support.
- nautilus12 6y agoThis is a better alternative to ES strictly for log analysis application. Not for search.
- lun4r 6y agohttps://vespa.ai/ https://vespa.ai/
- woof 6y agoApache Solr? Maybe not "better", but useful and with a Apache 2.0 license :)
- lovelearning 6y agoBetter in what ways?
- StreamBright 6y agoIn many ways. I could write a book about this but lets just say operational stability and reliability is the primary reasons.