Monitoring provides feedback at the application level and at the system level. “One of our customers is a social media company that uses Chronosphere to determine whether people are logging in and posting,” Dines remarked. “A change from the normal rate could be detected at this level before the issue is seen at the system level.” Monitoring can also produce actionable information. “If orders for a food delivery service spike and the number of drivers drops, then the company can call for additional drivers.” Other common metrics include shopping cart abandonment and revenue streams by product or by region.
Developers who are responsible for preventing downtime in cloud environments need a software tool that is intuitive and allows them to identify the source of the problem. “They are looking at an ocean of data, and the software should tell them where to start,” commented Dines. “Many existing proprietary tools were not designed for developers, but for IT or security staff. In the modern cloud computing era, the person who writes the code for business analytics, ecommerce, or social networking is also usually responsible for making sure the system is working correctly.”
The other issue that drives companies to explore new observability tools is the cost of the ones they are using. “Sometimes, observability costs reach 20%– 30% of their overall cloud computing plan, which is not sustainable,” Dines noted. “They should be around 10%, or less.” Companies using cloud-native architectures have reported that the tooling used to detect and fix incidents has increased costs and time to resolve the incidents; Chronosphere is designed to be a more cost-effective solution. For example, it ingests all data produced by infrastructure and applications and gives customers the control to transform the data into its most useful form, typically reducing data volume by 60%.
Searching the cloud
Given the volume and diversity of data that is likely to be stored in the cloud, having a strong search engine is an essential component of cloud implementations. Federated search—a search solution that accesses multiple repositories through one interface—means that users do not need to know the application in which information is stored. This feature has particular benefits in cloud ecosystems, which are often larger than on-prem storage systems and can have dozens of different applications.
Federated search in the cloud uses the same process as on-prem searching. According to Jeff Evernham, VP of product strategy at Sinequa, “Search in the cloud is identical to searching locally. The connectors to the information are different, but the functionality is the same.” There are differences, however, primarily in the underlying infrastructure. Most cloud implementations are multi-cloud, using a mix of public and private resources, and often also include legacy on-prem content.
Sinequa first optimized its cloud search technology for Azure to match its customer base, all of whom use Microsoft. “Each cloud has its own unique features and pricing structure, so if we optimize how we store and compute, we can bring costs down,” added Evernham. “In addition, making the search service more elastic is automated rather than manual. Servers can be added when there are more queries at certain times of the day, for example.”