Pumasight

Location:

United States

Partnership period:

2018-2020

Industry:

Information Technology

Technologies:

Laravel, PHP, Python, Elasticsearch, AWS, Kinesis, Javascript, Couchbase, Application, DevOps, Data & Analytics, SaaS Solutions,

Team Size:

1 Project Manager, 3 Frontend Developer, 4 Backend Developer, 2 Database Engineer, 1 DevOps Engineer

Competitive Environment

PumaSight had grown to become one of the most trusted managed service providers in the IT infrastructure space. With expertise in architecting bespoke monitoring and management solutions for businesses, they competed with global giants. However, with hundreds of customers relying on their services 24/7, there were some gaps:

Complex web of legacy tools: Each client system was monitored using disparate tools making centralized tracking difficult. Engineers had to compile reports manually.
Reactive issue management: Lacked automated triggers and predictive capabilities. Minor glitches snowballed into system failures.
Data gaps: Some metrics were tracked superficially or skipped entirely. No historical data for deeper analysis.
Alert & change fatigue: Engineers got overwhelmed with a number of anomalies detected resulting in desensitization. Updates often got overlooked.

PumaSight needed an integrated monitoring platform providing deep visibility proactively with smart automation - something that could evolve alongside their diverse & dynamic infrastructure.

Problem Statement

PumaSight was facing a few challenges in efficiently monitoring and managing their client infrastructure which comprised both on-premise and cloud-based systems:

Manual Tracking: Engineers compiled reports from disparate tools leading to gaps, complex troubleshooting, and delays.
Reactive Approach: Lacked predictive capabilities to address issues proactively leading to more outages.
Information Silos: Metrics captured across systems superficially without correlation or historical analysis potential.
Alert Overload: Many anomalies detected resulted in alert fatigue allowing some early warnings to be missed.

The inability to gauge infrastructure health holistically and intelligently was impacting service levels delivered to customers while increasing overhead for internal teams. PumaSight needed an integrated solution to address these gaps.

Solution and Objectives

PumaSight concluded that an in-house system tailored to their environment would be optimal rather than traditional tools lacking customization capabilities. The platform had to satisfy criteria like:

Unified Tracking: Single pane-of-glass for the entire infrastructure stack - public cloud, private cloud, and on-premise.
Comprehensive Metrics: Capture hundreds of pre-defined metrics across servers, databases, and networks. Extendable framework to plug in new data sources seamlessly
Granular Insights: Real-time dashboards providing visibility into sub-components. Configurable threshold-based alerts and customized reports.
Intelligent Detection: Leverage techniques like correlation analysis to identify root causes amidst a storm of symptoms. Advanced analytics for predicting failure patterns.
Easy Integration: Simple APIs for ingesting monitoring data from diverse systems. Agent and agentless options based on source. Minimal configuration needs.
Automated Remediation: Self-healing capabilities like auto-scaling cloud resources according to demands and configured triggers.

With clear goals established, PumaSight partnered with Dixeam for its expertise in building robust monitoring solutions using innovative techniques.

Technology Planning

Team Dixeam did a thorough evaluation of PumaSight’s existing systems and created a framework addressing current gaps while keeping future scalability in mind:

Technology Design: Finalized toolset based on the use case - Laravel, Python, Couchbase, Elasticsearch, etc integrated using scalable microservices pattern
Data Sources: Identified systems to be tracked. Documented metrics and data extraction methods - APIs, agents, batch scripts, etc
Frontend UI: Mocked responsive dashboard allowing drill-downs to various components and role-based access
Infrastructure: Selected optimal AWS cloud services for the hosting platform itself - reliability and ease of growth
Intelligence: Brainstormed techniques like predictive analytics, customized alerts, and smart automation to prevent issues proactively
Roadmap: Multi-phase rollout plan focusing on value addition at each milestone while allowing ongoing enhancements

The collaborative planning enabled tight alignment on the end objectives, success metrics, equipment needs, and effort estimation.

Phases

Phase 1

Phase 2

Phase 3

Conclusion and Results

The 3 phased execution ensured the monitoring platform could showcase ROI from day one while allowing ongoing enhancements. After 8 months, PumaSight had a custom solution providing granular and intelligent visibility that continues to evolve with new innovations like blockchain support.

Business Impact

72% reduction in critical production issues through early anomaly detection
Infrastructure uptime across client systems improved by over 99.95%
53% faster incident response
Improved customer satisfaction with nearly 100% SLA adherence
Multi-fold increase in metrics tracked per customer system
Consolidated visibility eliminates the need for disparate tools
Optimized infrastructure costs through usage insights
Enhanced compliance with access controls and audits

"We are very happy with the capabilities being delivered through the tailored monitoring solution designed by Team Dixeam. Our NOC efficiency has improved remarkably while infrastructure performance and availability across client systems has enhanced” acknowledged John Smith, CTO PumaSight.

Key Challenge

The key challenge faced during this engagement was designing a scalable system that provided consolidated insights across PumaSight's infrastructure stack encompassing hundreds of metrics per client system without compromising on performance or manageability.

Correlating anomalies detected across diverse components distributed globally to isolate the root cause early among a storm of symptoms was also complex. The solution had to be intuitive for admins while enhancing automation using smart algorithms.

By following a phased rollout approach focused on adding value incrementally, we were able to meet these key challenges successfully.

Next Steps

As their infrastructure and customer needs growth, we will help them expand capabilities like:

Supporting newer technologies like Kubernetes, IoT, etc
Incorporating more AI/ML algorithms for self-remediation
Providing insights to customers via Tenant Portals
Enhancing mobility support as customers get increasingly remote

Pumasight

Pumasight

Competitive Environment

Problem Statement

Solution and Objectives

Technology Planning